Full opinion text
MEMORANDUM OPINION AND ORDER SCHREIER, District Judge. TABLE OF CONTENTS I. Parties & Background.980 II. History of Districts 26 and 27 .980 III. 2000 Census Data and the Current Legislative Plan.982 IV. Section 2 of the Voting Rights Act.986 A. Sufficiently Large and Geographically Compact.987 B. Minority Political Cohesiveness .995 1. Dr. Cole’s Analysis.996 2. Dr. Zax’s Analysis.998 3. Reliability of Each Method.1001 4. Non-Statistical Evidence of Cohesiveness.1004 5. Partisanship and Low Voter Turnout .1008 C. Usual Defeat of Indian-Preferred Candidates .....1010 D. Totality of the Circumstances..1017 1. History of Discrimination.1018 a. Voting.1018 b. Discrimination in Representation.1023 c. Recent Electoral Processes.1023 d. Access to Polling Places.1026 e. 1975 Amendments to the Voting Rights Act.1027 f. Redistricting and Representation .1028 g. Other Evidence of Official Discrimination.1028 h. Unofficial Discrimination.1031 2. The Extent of Racially Polarized Voting .1034 3. Use of Voting Procedures for Discriminatory Purposes.1036 4. Access to Candidate Slating Process.1037 5. Socioeconomic Disparities.1037 6. Racial Appeals in the Political Process.1041 7. Indian Elected Officials.1042 8. Unresponsiveness.1043 9. Tenuousness.1047 10. Proportionality.1048 11. Indian Candidacies.1049 12. Voter Apathy and Low Turnout.1050 V. Remedy.1052 INTRODUCTION Plaintiffs contend that South Dakota’s 2001 legislative redistricting plan dilutes Indian voting strength by packing District 27 with a 90 percent supermajority of Indians, in violation of § 2 of the Voting Rights Act of 1965. This, plaintiffs contend, minimized the total number of districts in which Indians could select the candidate of their choice. Plaintiffs seek to create at least one additional single-member house district with a majority of Indians as a remedy. Defendants deny the allegations. After considering the evidence admitted during a nine-day court trial, the court determines by a preponderance of the evidence the following facts and conclusions of law. I. Parties & Background Plaintiffs Alfred Bone Shirt and Belva Black Lance are Indians, qualified electors, members of the Rosebud Sioux Tribe, and residents of Todd County, which is currently part of District 27. T.III p. 634, 688. Plaintiffs Bonnie High Bull and Germaine Moves Camp are Indians, qualified electors, and residents of Bennett and Jackson Counties, respectively, which are currently part of District 26. Joint Stipulations of Fact § 1 (Docket 267). Defendant Chris Nelson is South Dakota’s Secretary of State, Secretariat of the State Election Board, and the successor in office to the original defendant, Joyce Ha-zeltine. Complaint (Docket 1); Answer (Docket 23); T.VI p. 1513. Defendant South Dakota House of Representatives is one of two houses of South Dakota’s legislature. Defendant Matthew Michaels is the Speaker of the House of Representatives and the successor-in-office to Scott Eccarius, an original defendant in this case. Complaint (Docket 1); Answer (Docket 23); T.VIII p. 2215. Defendant South Dakota Senate is South Dakota’s other house of the legislature and defendant Arnold Brown is the President Pro Tempore of the South Dakota Senate. Complaint (Docket 1); Answer (Docket 23); T.VIII p. 2158. Plaintiffs filed suit on December 26, 2001, alleging that South Dakota’s 2001 legislative redistricting plan (“the Plan”) violates their rights under §§ 2 and 5 of the Voting Rights Act of 1965. Complaint (Docket 1). On January 29, 2002, a three-judge panel heard plaintiffs’ § 5 claim and held that defendants violated § 5 by failing to preclear the Plan. Only the § 2 claim now remains to be decided by this court. See Bone Shirt v. Hazeltine, 200 F.Supp.2d 1150 (D.S.D.2002). II. History of Districts 26 and 27 From 1973 to 1975, a task force analyzed Indian/State government relations to improve tribal and state relations in South Dakota. It consisted of nine tribal chairmen, two senators, two representatives, and five lay people. Thomas Short Bull, a member of the Oglala Sioux Tribe (OST), was the executive director. T.II p. 488-89; Ex. 268-270. As of 1970, there were 28 legislative districts; none were majority Indian, and no Indians had been elected under that plan. One of the task force’s reports discussed voting as it related to Indians in South Dakota. T.II p. 491-94; Ex. 18 p. 162-185; Ex. 267 p. 14. At trial, Short Bull testified that the task force concluded that the legislative district plan gerrymandered the Rosebud and Pine Ridge Reservations by “divid[ing it] into three legislative districts, effectively neutralizing the Indian vote in that area.” None of the districts were majority Indian. T.II p. 491; Ex. 18 p. 167; Ex. 269 p. 9. A 1974 report by the task force recommended that Shannon, Washabaugh, Todd, and Bennett Counties be combined into one legislative district, which would be a majority Indian district. T.II p. 495; Ex. 18 p. 186; Ex. 267 p. 25. The legislature never considered the plan. Short Bull testified that “the state representatives and senators felt it was a political hot potato.... [T]his was just too pro-Indian to take as an item of action.... [The legislature] always tried to get ... a compromise on things, but there really wasn’t much to compromise here; either you gerrymander it, or you try to have a district that has more Indian people in the legislative district.” The task force did propose eight other bills in 1974 that affected Indians, and the legislature passed seven of them. T.II. p. 497-506; Ex. 268 p. 43. In 1975, the legislature chose to not fund the task force and it was dissolved. A state Civil Rights Commission followed instead. T.II p. 506, 514. Following the 1980 census, the legislature drew a new redistricting plan. The state Civil Rights Commission recommended that the state should create a district in the area of the Rosebud and Pine Ridge Reservations. After the national Civil Rights Commission received the state commission’s report, the Department of Justice instructed South Dakota that it would not approve its reapportionment plan unless the state created a substantially Indian district. T.II p. 513-14. The redistricting plan that was enacted in 1981 had 35 single-member senate districts. Each district also elected two house members. District 28, which included Shannon and Todd Counties and half of Bennett County, became the first majority Indian district in South Dakota. SL 1881, ch. 14, § 2; 1983, ch. 10, § 2. In 1982, District 28 was 86 percent Indian. Short Bull was elected from District 28 in 1982 and became the first Indian state Senator. T.II. p. 488, 515-16. The court finds Short Bull’s testimony credible. His knowledge and experiences as the director of the task force and as the first Indian state senator in South Dakota make his testimony reliable, credible, and probative. Accordingly, the court accepts his testimony and gives it substantial weight. Following the 1990 census, the legislature again engaged in re-districting. Ex. 869; Ex. 870. Meetings on the issue were held in June 1991 at Sinte Gleska College in Mission, South Dakota, the Bennett County Courthouse, and Oglala Lakota College (OLC) in Kyle, South Dakota. Ex. 869 p. 1. Testimony at the meetings expressed concern about diluting the Lakota (Indian) vote and about drawing districts “to maximize the opportunity for Lakota people to elect a Lakota to represent them.” People testified about concerns regarding the accuracy of the census results. They believed the reservations had a “significantly greater population than the Census indicated.” Although some testimony urged the legislature to retain the current boundaries, there was also testimony that if a census recount evidenced a strong population increase, “two Indian districts would be preferable.” Ex. 869 p. 1-2. The re-districting plan adopted in 1991 retained a district consisting of Todd and Shannon Counties and a part of Bennett County, except the city of Martin. 1st SS 1999, ch. 1, § 3. This district was renumbered District 27. Id. The redistricting plan did not create a second district containing a majority of Indians. Ex. 869; Ex. 870. During a second special session in 1991, the legislature created two single-member house districts, namely Districts 28A and 28B, which encompass part of the Standing Rock Reservation and all of the Cheyenne River Indian Reservation. According to the statute, the districts were created for the purpose of protecting minority voting rights. 2nd SS 1991, ch. 1. III. 2000 Census Data and the Current Legislative Plan South Dakota’s Constitution requires reapportionment of its membership every ten years after 1991. S.D. Const, art. Ill, § 5. The legislature convened on October 23-24, 2001, for the purpose of redistricting. Ex. 401 Tab 2; Ex. 402. When drafting the 2001 Plan, the legislature relied on the Census 2000 Redistricting Data Summary File produced by the United States Census Bureau. This was the first federal Census that allowed respondents to identify themselves with more than one racial group. Consequently, there are several ways to assign people to racial groups for the purpose of redistricting. T.I p. 258, 260. The single-race method classifies only those who identify themselves exclusively as a single racial group within that category. The multiple-race method assigns each multiple race response to every category with which the person identifies. Black Political Task Force v. Galvin, 300 F.Supp.2d 291, 301 (D.Mass.2004). The dual-race method assigns any multiple-race response that includes white and one of the other five race categories to the minority race listed in the response. Guidance Concerning Redistricting and Retrogression Under Section 5 of the Voting Rights Act, 66 F.R.D. 5412-01, 5414 (2001). Because this “case involves an examination of only one minority group’s effective exercise of the electoral franchise, ... it is proper to look at all individuals who identify themselves as [Indian.]” Georgia v. Ashcroft, 539 U.S. 461, 123 S.Ct. 2498, 2507-8 n. 1, 156 L.Ed.2d 428 (2003). This court will, therefore, consider all people who self-identify as Indian, whether single, dual, or multiple race. See id.; T.I p. 260-263. According to the 2000 census, the total population of the state of South Dakota is 754,844, with an Indian population of 67,-990. Thus, Indians constitute 9.05 percent of the total population and 6.79 percent of the voting age population (VAP). Ex. 358 p. 6. sub-ex.l. There are nine federally recognized tribes that have land in South Dakota. All of “Indian country” is located in rural areas. The total population of South Dakota’s “Indian country” is 59,366, of whom 42,460 are Indians. Ex. 358 p. 7-8. Nearly two-thirds (62.55 percent) of South Dakota’s Indian population lives in “Indian Country.” Eight predominantly rural counties are Indian majority: Shannon, Bennett, Todd, Mellette, Dewey, Ziebach, Corson, and Buffalo. Ex. 358, Ex. 2. The redistricting committee consisted of 15 members. Senator Arnold Brown was the committee co-chair. Two Indian legislators were members of the committee, namely Representative Paul Valandra (member of the Rosebud tribe) and Senator Dick Hagen (member of the Oglala Sioux tribe). Both are elected officials from District 27. Representative James Bradford, also from District 27, requested placement on the committee but was told there was not enough room for another member. T.VIII p. 2164; T.IV p. 1080; Ex. 401. The 2000 redistricting committee held six meetings, four in Pierre, one at Sinte Gleska College on the Rosebud Reservation, and one at Wolf Creek school on the Pine Ridge Reservation. Legislative members from Districts 26 and 27 encouraged the holding of meetings at Sinte Gles-ka and Wolf Creek to provide an opportunity for local input. Senator Hagen and Representative Valandra were in charge of establishing the meeting dates and advertising. Both meetings were poorly attended. Six people attended the Wolf Creek meeting, and only one person spoke. T.VIII p. 2168-71, 2242; T.IV p. 1101. Representative Bradford testified that although Senator Hagen publicized the occurrence of a meeting, he did not adequately advertise the location or time. Originally the meeting was planned to be held at Billy Mills Hall which is located in the town of Pine Ridge; however, Senator Hagen failed to reserve the hall. Representative Bradford offered Wolf Creek school for the meeting, which is located five miles outside of the town of Pine Ridge. Representative Bradford attributes the low attendance at the meeting to its inconvenient and unadvertised location. T.IV p. 1081-82. Representative Valandra set up the meeting at Sinte Gleska College. Representative Bradford testified that Indians were “very, very sensitive to the atmosphere.” The formality of the room and the committee’s attire made Indians feel like they were inadequate “to appear before a committee of that magnitude.” T.IV p. 1100-01. Following the redistricting committee hearings, the legislature took up the issue of redistricting at a special session held on October 23 and 24, 2001. Like previous plans, the 2001 Plan divided South Dakota into 35 legislative districts, each of which elects one member to the Senate and two members at large to the House of Representatives, except for District 28. District 28 is divided into two single-member districts, Districts 28A and District 28B, which each elect one member to the House of Representatives. SDCL 2-2-34. Under the Plan, District 27 includes all of Shannon and Todd Counties, and Precinct 27 of Bennett County, which is the southern portion of Bennett County. District 26 borders District 27 and includes all of Haakon, Jackson, Jones, Lyman, Mellette and Tripp Counties, and Precinct 26 of Bennett County, which is the northern portion of the county. SDCL 2-2-34. During the house floor discussion, Representative Bradford introduced an amendment to provide for a new configuration of Districts 26 and 27, which would have created a majority-Indian Senate District 27 consisting of Haakon, Jackson, and Shannon Counties and a portion of Bennett and Todd Counties. Senate District 26 would have consisted of Jones, Mellette, Lyman, and Tripp Counties. Two single-member house districts would have been created for District 26, wherein majority-Indian District 26A would have consisted of Mellette County and a portion of Bennett, Jones, and Todd Counties. District 26B would have consisted of the remainder of District 26 and would have been a majority-white district. The amendment was defeated by a vote of 51 to 16. Ex. 402, p. 18 of House Journal. The bill was passed by both houses and the governor signed the bill into law on November 1, 2001. Ex. 402 p. 37; SDCL 2-2-34. A map of the adopted plan follows as Figure 1. Every district in the 2001 Plan is majority white except Districts 27 and 28A, which are majority Indian. According to the 2000 Census, Indians make up approximately 90 percent of District 27’s total population and 86 percent of the VAP. Indians make up approximately 30 percent of District 26’s total population and 23 percent of the VAP. Ex. 358 p. 14; Ex. 369. Part of the Pine Ridge Reservation and all of the Rosebud Reservation are located within District 27. This district elects one senator and two representatives to the legislature. Part of the Standing Rock Reservation and all of the Cheyenne River Reservation are within House District 28A. This district elects one representative and is half of a district that elects one senator to the legislature. Ex. 358 p. 14 The court finds the testimony of Representative Bradford to be highly credible. He grew up in Parmalee, South Dakota, on the Rosebud Reservation and lived there for 15 years as an adult. He graduated from Sinte Gleska College, which is also located on the Rosebud Reservation. He has lived in Mission, South Dakota, on the Rosebud Reservation and near Martin, South Dakota, in Bennett County for 15 years. His wife is from Pine Ridge, and they have lived there since 1978. His mother was part Indian, and although neither his mother nor he are enrolled members, he is eligible for tribal benefits such as Indian Health Service and education and he has always identified himself as an Indian person. He has been a representative in the South Dakota legislature since 2000. T.IV p. 1059-64. The court finds that his experiences and knowledge about Indian issues make his testimony probative. The court, therefore, accepts his testimony in its entirety and gives it great weight. Representative Valandra and Senator Hagen preferred keeping the districts in the current form. Representative Matthew Michaels, a member of the redistricting committee and Speaker of the House of Representatives, testified that Senator Hagen stated that “he had not heard an outcry to change [the districts]” and that people who testified at the meetings on the Rosebud and Pine Ridge Reservations wanted to “leave the status quo.” T.VIII p. 2173, 2180-81, 2231-32. Representative Bradford proposed an alternative map that would have increased the number of Indian-majority districts. T.VIII p. 2180-81. Although no evidence refutes that Representative Valandra and Senator Hagen told others on the redistricting committee that they would prefer the district boundaries to be left unchanged, the court gives little weight to this evidence because as incumbents, they had a vested interest in resisting change to their district’s boundaries. The limited public testimony at the Sinte Gleska and Wolf Creek meetings also does not necessarily represent the attitudes of Indians in the affected districts. The low attendance at these meetings detracts from the probative value of the evidence. Furthermore, there is evidence that directly refutes this evidence. During the special session in October 2001, Representative Valandra voted in favor of an amendment proposed by Representative Bradford that would have changed the boundary lines of Districts 26 and 27. Ex. 402 p. 19. Additionally, lay testimony at trial undermines the claim that Indians wanted to keep Districts 26 and 27 unchanged. Elsie Meeks testified that the current configuration of Districts 26 and 27 is unfair because it “segregates Indians.” T.V p. 1376. Lyla Young described the current districts as unfair because they “dumped us all in this bloc,” which gives Indians less choice. T.III p. 676. Short Bull stated that District 27 was unfair because there were “just too many Indians in that legislative district,” which dilutes their vote. T.II p. 518. Defendants point to evidence that it was reasonable for legislators to rely on the statements by Senator Hagen and Representative Valandra. T.IV p. 1094. While their reliance may not be unreasonable, the court does not find this evidence probative of whether the Plan violates § 2. Regardless of Senator Hagen’s and Representative Valandra’s statements, legislators were obligated to draw district lines that comply with § 2. This evidence, therefore, has little probative value. Defendants also argue that ACLU representative Jennifer Ring testified at a redistricting meeting that the “Sagebrush/Cocklebur Plan” was her favorite. This plan is virtually the same as the current plan adopted by the legislature. Ex. 401 p. 5, Tab 10 p. 14. Ring testified at trial, however, that she did not choose a favorite and that she believed District 27 was a “packing plan.” T.III p. 767, 811-12. Furthermore, whether or not Ring favored a particular plan is irrelevant to determining whether Districts 26 and 27 comply with § 2 of the Voting Rights Act. Thus, the court does not give this evidence great weight. IV. Section 2 of the Voting Rights Act 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 U.S.C. §§ 1973(a), 1973b(f)(2); Thornburg v. Gingles, 478 U.S. 30, 44, 106 S.Ct. 2752, 2763, 92 L.Ed.2d 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 U.S.C. § 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 U.S. 997,1007,114 S.Ct. 2647, 2655, 129 L.Ed.2d 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. Quitter, 507 U.S. 146, 154, 113 S.Ct. 1149, 1155, 122 L.Ed.2d 500 (1993). 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 U.S. at 51, 106 S.Ct. 2752. 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 U.S.C. § 1973(b). A. Sufficiently Large and Geographically Compact 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 U.S. at 50,106 S.Ct. 2752. Small and dispersed minority groups undermine the ability to create a district that would remedy the grievance. Sanchez v. Colorado, 97 F.3d 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 F.3d 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. Id. at 50 n. 17,106 S.Ct. 2752. This 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 F.3d 92, 95 (5th Cir.1994) (Clark I). The constitution does not require regularity of district shape. Bush v. Vera, 517 U.S. 952, 963, 116 S.Ct. 1941, 1953, 135 L.Ed.2d 248 (1996); Sanchez, 97 F.3d at 1312. Plaintiffs, therefore, need only propose a plan that demonstrates the possibility of a majority-minority district. Houston v. Lafayette County, Miss., 56 F.3d 606, 611 (5th Cir.1995). 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 F.3d at 1312. Plaintiffs introduced into evidence five illustrative redistricting plans which were drawn by their demographic expert, William Cooper. Each plan creates at least one additional majority-Indian house district. Ex. 358 p. 16, sub. ex. 9-13; T.I p. 270-71. Illustrative Plan A creates three majority-Indian districts, including two majority-Indian senate districts and one majority-Indian single-member house district. It divides one of the majority-Indian senate districts into two single-member house districts, one of which is majority-Indian. Thus, six legislative seats are majority-Indian. The following table summarizes the total population and VAP. Ex. 358 p. 16-17, sub. ex. 9; Ex. 369. _Single Race Dual Race Multi-Race District 26 Senate (VAP)_48.57_50.15_50.20_ _(total population)_57.03_58.96_59.01_ District 26A single-member House (VAP) 67.58_69.81_69,88 _(total population)_74.52_77.01_77.08 District 27 Senate (VAP)_70^0_7L41_71.42 _(total population)_76.67_77.91_77.92 District 28A single-member House (VAP) 67.00_67.94_67.98 (total population)74.13_75.26_75.32 Illustrative Plan B creates two majority-Indian senate districts and divides one of the majority-Indian senate districts into two single-member house districts, both of which are majority-Indian. A total of six legislative seats in Illustrative Plan B are majority-Indian. The table summarizes the total population and VAP. Ex. 358 p. 17-18 sub-ex. 10; Ex. 369. Single Race Dual Race Multi-Race District 26 Senate (VAP) 61.10 62.53 62.57 (total population) 68.57 70.22 70.28 District 26A single-member House (VAP) 69.10 69.94 69.99 (total population) 75.65 76.70 76.76 District 26B single-member House (VAP) 53.20 55.20 55.24 (total population) 61.35 63.61 3.68 District 27 Senate (VAP) 84.81 86.04 86.07 (total population) 88.66 89.91 89.9 Illustrative Plan C creates one majority-Indian senate district and two majority-Indian single-member house districts. A total of five legislative seats are majority-Indian. The table summarizes the total population and VAP. Ex. 358 p. 18-19 sub-ex. 11; Ex. 369. Single Race Dual Race Multi-Race District 26A single-member House (VAP) 65.70 67.8 67.89 (total population) 73.08 75.62 75.69 District 27 Senate (VAP) 9.21 70.46 70.47 (total population) 75.81 77.09 77.11 District 28 Senate (VAP) 68.3 69.25 69.29 (total population) 75.03 76.04 76.10 Illustrative Plan D creates one majority-Indian senate district and two majority-Indian single-member house districts. A total of five legislative seats are majority-Indian. The table summarizes the total population and VAP. Ex. 358 p. 19-20 sub-ex. 12; Ex. 369. Single Race Dual Race Multi-Race District 26A single-member House (VAP) 65.60 66.65 66.70 (total population) 73.06 74.21 74.25 District 27 Senate (VAP) 67.19 68.96 3.98 (total population) 74.05 75.Í 76.02 District 28A single-member House (VAP) 68.38 69.25 69.29 (total population) 75.03 76.04 76.10 Illustrative Plan E creates one majority-Indian senate district and two majority-Indian single-member house districts. A total of five legislative seats are majority Indian. The table summarizes the total population and VAP. Ex. 358 p. 20-21 sub-ex. 13; Ex. 369. _Single Race Dual Race Multi-Race District 26A single-member (VAP)_73.10_74.31_74.36_ _(total population)_79.39_80.84_80.88_ District 27 Senate (VAP)_63.87_65154_65.56 _(total population)_71.16_72,98_73.00_ District 28A single-member (VAP)_68.38_69.25_69.29_ (total population)75.03_76.04_76.10 Cooper’s plans present various alternatives, each of which creates at least one additional majority-minority legislative district in South Dakota while adhering to traditional redistricting principles. For instance, Plan E largely follows existing county and reservation boundaries. For illustrative purposes, Plan E is shown here as Figure 2. Plans A and B, while splitting some counties, avoid splitting precincts. The proposed districts are also compact and relatively contiguous. T.I p. 271; Ex. 358 sub-exs. 9-13. The fact that Cooper’s socio-economic data reports only information relating to single-race Indians does undermine the credibility of his report. This section of the report was derived from census data that only provided information for single-race Indians. As noted previously, there is not a significant difference between the number of people who self-identified themselves as single, dual, or multiple race Indians. Although less comprehensive, Cooper’s use of single-race Indian data does not discredit his redistricting maps or his findings. Furthermore, even though Cooper has not taught at a college, written for a journal, and is not a sociologist, political scientist, economist, or econometrician, he is nonetheless credible and qualified as an expert to draw redistricting maps. Neither his testimony nor his report require expertise in these social sciences for purposes of providing reliable testimony about alternative redistricting plans for South Dakota. He need not be an expert in anthropology, Sioux culture and history, or South Dakota history to reliably report on redistricting options in South Dakota. He can reliably base his analysis and conclusions on his experience in South Dakota and his knowledge of redistricting. T.II p. 306-14. Furthermore, the fact that Cooper did not analyze what effect, if any, discrimination had on producing a relatively depressed economy and less prosperous population of Indians than non-Indians in Districts 26 and 27 does not undermine his analysis, conclusions, or illustrative plans. In fact, this issue is not in real dispute because even defendants’ expert, Michael Lawson, acknowledged that past discrimination has contributed to present day depressed socioeconomic status among Indians. T.II p. 333-85; T.VIII p. 2116. Defendants argue that plaintiffs did not satisfy the first Gingles factor because they failed to propose a complete, narrowly tailored remedy that contains constitutionally workable districts. Plaintiffs need not propose a complete remedy at this stage. See Gingles, 478 U.S. at 50 n. 17, 106 S.Ct. 2752 (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 simply demonstrate the feasibility of drawing a majority-minority district and are not cast in stone. Houston, 56 F.3d at 611. See also Dickinson v. Indiana State Election Bd., 933 F.2d 497, 503 (7th Cir.1991) (completeness of remedy considered at the remedial stage of litigation). “If a § 2 violation is found, the [state] will be given the first opportunity to develop a remedial plan.” Id. See also Clark v. Calhoun County, Miss., 88 F.3d 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 plaintiffs”); Sanchez, 97 F.3d at 1315 (“drawing the necessary district is not [plaintiffs’] onus because the State must be given the first opportunity to fashion a remedy”). Defendants also argue that plaintiffs have failed to satisfy this factor because any proposed remedy requires above a supermajority of 65 percent Indian YAP. The court disagrees. First, the law does not definitively require establishing super-majority districts. Gingles, 478 U.S. at 50, 106 S.Ct. 2752 (requiring proof of a majority); Valdespino v. Alamo Heights Ind. Sch. Dist., 168 F.3d 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 F.2d 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). Second, any necessary supermajority is required only at the remedial stage of litigation. Indeed, the 65 percent guideline is a general remedial goal and “is irrelevant to the first part of the Thorn-burg tripartite threshold test for liability.” Magnolia Bar Ass’n v. Lee, 793 F.Supp. 1386, 1397 (S.D.Miss.1992). In Dickinson, 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 F.2d at 503 (citations omitted). Because the current case is at the liability stage of the proceedings, the court concludes that proof of a supermajority is not required. Defendants rely on Jeffers v. Tucker to support their contention that a superma-jority is necessary. Jeffers notes that in fashioning remedies for Voting Rights Act violations, the creation of districts with bare majorities is not enough for a complete remedy. In order to compensate for historically low rates of voter registration and turnout, “minorities must have something more than a mere majority even of voting age population in order to have a reasonable opportunity to elect a representative of their choice.” 847 F.Supp. 655, 660 (E.D.Ark.1994). Defendants’ argument fails for several reasons. First, the current stage of the litigation is determining liability, not fashioning a remedy. Second, plaintiffs have proposed five plans which maintain at least a 65 percent Indian VAP in District 27 and create a new single member house district with at least a 65 percent Indian VAP in District 26A. ■ Third, a 65 percent supermajority does not necessarily apply. In James v. City of Sarasota, Fla., the district judge referenced the 65 percent guideline. 611 F.Supp. 25, 28 n. 3 (M.D.Fla.1985). The court later attached a letter from the chief of the civil rights division of the Justice Department to the opinion: There is no 65 percent threshold population figure.... Rather, our responsibility is to determine whether the number of districts in which minority voters will have a fair opportunity for electing representatives of their choice (whatever the percentage of minority voters in each district) has been decreased by the proposed redistricting plan submitted under Section 5.... In this context, the Section 5 inquiry is much less concerned with discrete percentages of minority voters in reconfigured districts ... [We do not attach] particular significance to a 65 percent population figure, and no attempt is made to add arbitrarily increments of five percentage points each to compensate for age, registration and turn-out differences.... Each Section 5 submission must, of course, be evaluated in light of the particular factual circumstances ... not on the basis of some preordained population percentage. Id. at 32-33. Even though DOJ’s letter in James referenced a § 5 violation, the court finds it persuasive on the issue of whether the law requires proof of a particular percentage of minority voters in a proposed district. The law does not require a discrete number. See Magnolia. Bar Ass’n, 793 F.Supp. at 1397 (“The 65 percent guideline, therefore, is a general remedial goal in Voting Rights Act cases that is irrelevant to the first part of the Thornburg tripartite threshold test for liability.”); Neal v. Coleburn, 689 F.Supp. 1426, 1438 (E.D.Va.1988) (“Contrary to defendants’ contention, the general 65 percent 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 percent standard is a flexible and practical guideline to consider in fashioning relief for a § 2 violation.”). Accordingly, the court finds that no supermajority is required at this stage of the litigation, and even if the law did so require, plaintiffs have proposed several remedies that satisfy a 65 percent supermajority standard. While defendants contend that the su-permajority should be higher than 65 percent, defendants have not identified any case law which requires a supermajority higher than 65 percent. Gingles requires only that a plaintiff show that minority voters “possess the potential to elect representatives.” Gingles, 478 U.S. at 50 n. 17, 106 S.Ct. 2752. Section 2 is not a guarantee that minority voters will elect the representative of their choice. Therefore, the court is not persuaded that a supermajority greater than 65 percent is required. Additionally, defendants contend that plaintiffs’ plans are race based, are impermissible gerrymanders, and fail to account for communities of interest. In Jeffers, 847 F.Supp. at 655, Chief Judge Arnold found that any plan “would have to be consistent with the spirit of Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993).” In Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993), the Court held that an equal protection claim that triggers strict scrutiny is raised when a voting scheme is “so extremely irregular on its face that it rationally can be viewed only as an effort to segregate the races for purposes of voting, without regard for traditional districting principles and without sufficiently compelling justification.” Id. at 2824. Here, plaintiffs have proposed several redistricting plans that do not appear any more irregular or gerrymandered on their face than the plan that the legislature enacted. Furthermore, “drawing racial distinctions is permissible where a governmental body is pursuing a ‘compelling state interest.’ ” Shaw v. Hunt, 517 U.S. 899, 908, 116 S.Ct. 1894, 1902, 135 L.Ed.2d 207 (1996) (Shaw II). The Supreme Court has assumed without deciding that compliance with the results test of § 2 of the Voting Rights Act (VRA) is a compelling state interest. See Shaw v. Hunt, 116 S.Ct. at 1905; Miller v. Johnson, 515 U.S. 900, 920-21, 115 S.Ct. 2475, 2490-91, 132 L.Ed.2d 762 (1995); Bush v. Vera, 517 U.S. at 977, 116 S.Ct. 1941. Justice O’Connor in a separate concurrence found that compliance with VRA in fact is a compelling state interest. Vera, 116 S.Ct. at 1968. A state 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. Additionally, the proposed plans are not so irregular on their face that they appear to be solely an effort to segregate races for the purpose of voting. In fact, the proposed plans are no more irregular than the current districting plan. Indeed, illustrative plans C, D, and E largely coincide with the current plan. Ex. 358 p. 19 sub-exs. 9-13; T.I p. 271. Plaintiffs’ plans consider race to the extent necessary to determine whether additional minority-majority districts are possible but not any more than reasonably necessary. Accordingly, the plans do not impermissibly subordinate race-neutral districting principles to racial considerations. T.II p. 281-82; Ex. 358 sub-exs. 9-13. Furthermore, Figure 3, which is a South Dakota map based on Indian population by county, as shown here, reflects it is not difficult to draw an additional majority-minority legislative seat because the Indian population is concentrated in the area from the Crow Creek and Lower Brule Reservations toward the Pine Ridge and Rosebud Reservations. The areas between these reservations are also sparsely populated. T.I p. 271. In response to defendants’ claim that each plan fails to incorporate important community values and traditional redistricting principles, plaintiffs point to Cooper’s testimony as evidence that he adhered to South Dakota’s traditional redistricting principles when drafting the five plans. Traditional redistricting principles as recognized by the South Dakota state legislature include population equality, protecting community of interests through compact and contiguous districts, respect for geographical and political boundaries, and protection of minority voting rights. SDCL 2-2-32. Similar principles have been recognized as traditional redistricting principles by the United States Supreme Court. See, e.g., Miller, 515 U.S. at 916, 115 S.Ct. 2475 (identifying “respect for political subdivisions and communities defined by actual shared interests” as traditional districting principle); Shaw v. Reno, 509 U.S. at 651-52, 113 S.Ct. 2816 (identifying population equality as a traditional districting principle); Miller, 515 U.S. at 916, 115 S.Ct. 2475 (identifying compactness and contiguity as traditional districting principle). The court accepts Cooper’s explanation that he applied traditional districting principles. A review of the plans reveals that the plans attempt to keep county and precinct boundaries intact, adhere to traditional geographic boundaries, consider current and historical reservation boundaries, create districts with the population equality, create compact and contiguous districts, and make minimal changes to the 1991 Plan. T.II p. 281, 289-91, 339. Defendants stressed other nuances about South Dakota history and culture, including differences between East and West River and differences between farmers and ranchers, to discredit Cooper’s redistricting plans. Defendants contend Cooper’s plans combine East and West River counties into the same district and put farmers and ranchers into the same district. The court notes, however, that the 2001 Plan that was adopted by the legislature has two districts, 24 and 21, that include counties from both East and West River and several districts combine counties occupied by both farmers and ranchers. Thus, it appears that such concerns are not significant enough to discredit Cooper’s five plans. Furthermore, if the court finds a § 2 violation, defendants will have the first opportunity to draft a remedial plan and will have the opportunity to consider any differences between East and West River they deem relevant to redistricting. See Sanchez, 97 F.3d at 1315. The court finds that Cooper’s testimony is credible and probative. Cooper is a geographic information system consultant who works with census data to prepare maps that show demographic patterns or trends. He has extensive experience in drawing redistricting maps, having drafted “thousands and thousands” of maps. Cooper has prepared such maps since 1986 for approximately 475 jurisdictions to promote compliance with the Voting Rights Act. He has proposed redistricting plans for § 2 litigation for 18 states. He has also analyzed the 2000 census data for 100 local jurisdictions in 15 states and has developed plans for 8 states. He has testified as an expert witness in approximately 24 voting rights cases in 10 different states and has been involved in 19 other cases. T.I p. 241^4; Ex. 358 p. 1-3. Since 1999, Cooper has worked with various Indian communities in South Dakota relating to redistricting. In addition, he has researched socio-economic factors affecting Indians and voting rights. The court further finds that Cooper employs reliable methods, including a software mapping program that he has used extensively for 13 years. T.I p. 244-46. For the reasons previously detailed herein and for the reasons stated in this court’s order dated December 31, 2003, the court finds Cooper’s testimony credible and his methods reliable. The court gives significant weight to the analysis and findings contained in Cooper’s testimony and report. Defendants’ Objections to Plaintiffs’ Proposed Findings of Fact and Conclusions of Law (Docket 319); Ex. 358; Court Order 12/31/03 (Docket 204). In addition to the maps proposed by Cooper, there is additional evidence in the record that supports the conclusion that a permissible remedy in the context of the challenged system can be fashioned. During the 2001 redistricting process, legislative research staff members drafted several plans that added at least one additional majority-minority seat, including: the “Little Wound” plan, the “Spotted Tail” plan, and the “Hollow Horn Bear” plan. All committee members received copies of these maps. Ex. 401 Tab 7; T.V p. 1416. Although several of the proposed legislative plans had only a simple majority of Indians and no VAP majority of Indians, the population trends according to the 2000 census indicate that the Indian population in the counties at issue continues to grow rapidly while the white population declines. Ex. 401 Tab 7. Furthermore, the additional district created in the “Hollow Horn Bear” plan contained a greater than 65 percent VAP. After considering all the evidence, including Cooper’s five illustrative plans and the plans drafted by the legislative research staff, 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. B. Minority Political Cohesiveness The second Gingles factor requires plaintiffs to show that the minority group is politically cohesive. Gingles, 478 U.S. at 51, 106 S.Ct. 2752. 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 F.2d 1240, 1244 (5th Cir.1988). See also Ruiz v. City of Santa Maria, 160 F.3d 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). Proving this factor typically requires statistical evaluation of elections. Campos, 840 F.2d 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 U.S. at 57 n. 25, 106 S.Ct. 2752. Courts have relied on various statistical methods. See, e.g., Houston, 56 F.3d at 611 (use of bivariate ecological regression and extreme case analysis); Clark, 88 F.3d at 1397 (expert employed regression and homogenous precinct analysis); Sanchez, 97 F.3d at 1317-18 (court considered both ecological and multivariate regression analysis). This court will examine each method used by the experts in this case. 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. 359 p. 4, 7-9. A second technique is bivariate ecological regression analysis (BERA). Under this technique, precinct-by-precinet 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. 359 p. 4-5. A third technique is ecological inference (El), 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. T.VII p. 1786-87. 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 U.S. at 80-82, 106 S.Ct. 2752 (relying exclusively on interracial legislative contests to determine whether a legislative redistricting plan diluted the black vote); United States v. Blaine County, Mont., 368 F.3d 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 then exogenous elections. See Sanchez, 97 F.3d 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 F.3d 973, 990 (1st Cir.1995) (recent elections more probative); Meek v. Metropolitan Dade County, Fla., 985 F.2d 1471, 1482-83 (11th Cir.1993). 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 F.3d 600, 605-06 (4th Cir.1996) (minority preferred candidate may be white). Thus, the court will not limit its consideration to interracial elections. T.IV p. 908; T.VII p. 1829-31. 1. Dr. Cole’s Analysis Plaintiffs’ expert, Dr. Steven Cole, used HPA and BERA. Ex. 359 p. 4. He applied the dual-race method of identifying Indians. Dr. Cole examined all state legislative elections in the current Districts 26 and 27 and several interracial elections for statewide and county offices from 1986 to 2002. He aggregated his data across both districts for statewide offices. Ex. 359 p. 3. Dr. Cole generated tables to demonstrate his results. In the first table, using BERA, Dr. Cole used a single regression analysis, which examines the relationship between the precinct’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.II p. 383-386; Ex. 359 p. 26-27. 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.II p. 386; Ex. 359 p. 26-27. 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. T.II p. 387; Ex. 359 p. 26-27. In Table 2, Dr. Cole employed HPA. For this analysis, he relied on virtually all-white or all-Indian precincts, generally over 90 percent. Precincts with 90 percent or more non-Indian VAP provide an estimate of white crossover voting while precincts with 90 percent of more Indian VAP measure Indian cohesion. T.II p. 388; Ex. 359 p. 28-29. When drawing conclusions about cohesion, Dr. Cole averaged estimates of minority support across elections and relied on the overall pattern of results. Ex. 359 p. 15; T.II p. 389-90. When defining “cohesion,” Dr. Cole does not employ a strict numerical threshold. Rather, according to Dr. Cole, he measures political cohesiveness on a continuum, “starting slightly above 50 percent and going all the way up to 100 percent. There is no absolute cutoff that 59 percent is not cohesion and 60 percent is cohesion. ... It’s the degree to which people stick together is the measure of cohesiveness. There is no cut point.” T.II p. 412. 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 “cut point.” T.II p. 481-8. 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. 359 p. 6. Dr. Cole analyzed five interracial elections for the state senate: the 1998 general election for Districts 26 and 27, the 1994 general election for District 27, and the 1988 and 1990 general elections for District 28. Ex. 359 p. 12. He found the average estimate of Indian political cohesion in these races to be 83 percent, which is “highly politically cohesive.” T.II p. 389-90; Ex. 359 p. 15, 26-27. He analyzed nine interracial elections for the state house: the 1994 and 2000 Democratic primary election, and the 1992, 1994, 1996, 1998, and 2000 general elections for District 27. He also analyzed the 1986 and 1990 general elections for District 28. Ex. 359 p. 19-20. Using BERA, Dr. Cole estimated the average level of Indian political cohesion at 77 percent. Although the percentages varied slightly, Dr. Cole’s BERA analysis was largely consistent with his HPA analysis. Ex. 359 p. 40-46. Dr. Cole analyzed eight elections for the state senate which involved only non-Indian candidates: the 1992, 1996, and 2002 general elections for District 26. He also analyzed the 1986 and 1990 general elections for Districts 27 and 29 and the 1988 general election for District 29. Ex. 359 p. 16. Using BERA, Dr. Cole estimated the average level of Indian political cohesion was 80 percent. This was consistent with his HPA analysis. Ex. 359 p. 17, 30-35. Dr. Cole analyzed eight elections for the state house involving only non-Indian candidates: the 1992, 1994, 1996, 1998, 2000 and 2002 general elections for District 26, the 1988 general election in District 29, and the 1986 general election in District 29. Ex. 359 p. 22. Dr. Cole found that using BERA, the average estimate of Indian political cohesion for the top preferred candidate in these races was 73 percent. This analysis was largely consistent with his HPA analysis. Ex. 359 p. 47-54. Dr. Cole analyzed six elections for the state senate involving only Indian candidates: the 1996 and 2002 Democratic primary election, the 1992, 1996, 2000, and 2002 general election for District 27. Ex. 359 p. 17-18. Using BERA, the average Indian cohesion for the six contests was 89 percent. Dr. Cole’s HPA analyses were largely consistent with his BERA analyses. Ex. 359 p. 18, 36-39. Dr. Cole analyzed one election for the state house involving only Indian candidates: the 2002 general election for District 27. Ex. 359 p. 22. Using BERA, Dr. Cole estimated Indian political cohesion for the top preferred candidate in this race at 80 percent. Dr. Cole’s BERA analyses were consistent with his HPA analyses. Ex. 359 p. 22, 55-56. Dr. Cole also analyzed eight exogenous interracial elections: the 2002 nonpartisan primary election for Bennett County School Board, the 2002 Democratic primary for Bennett County Commissioner at Large, the 2002 general elections for Attorney General, Bennett County Commissioner at Large, Register of Deeds and County Coroner, and the 1998 general election for Governor and Lieutenant Governor. Ex. 359 p. 23-24. Using BERA, Dr. Cole determined that Indian political cohesion for the top preferred candidate in these races was 87 percent. His BERA analyses were consistent with his HPA analyses. Ex. 359 p. 57-62. Across the 22 interracial elections that Dr. Cole analyzed with BERA, the average estimate of Indian political cohesion for the top preferred candidates was 82 percent. Ex. 359 p. 26-27, 40-42, 57-29. Across the 37 endogenous elections that Dr. Cole analyzed with BERA, the average estimate of Indian political cohesion for the top preferred candidate was 80 percent. Ex. 359 p. 26-27, 30-32, 36-37, 40-42, 47-50, 55. Across the 12 elections in 2002 that Dr. Cole analyzed with BERA, the average estimate of Indian political cohesion for the top preferred candidate was 84 percent. Ex. 359 p. 30, 36, 47, 55, 57-59. Across the 45 elections Dr. Cole analyzed using BERA, the average estimate of Indian political cohesion for the top preferred candidate was 81 percent. Ex. 359 p. 26-27, 30-32, 36-37, 40-42, 47-50, 55, 57-59. His results demonstrate high political cohesion among Indians in Districts 26 and 27. Ex. 359. The court finds that Dr. Cole qualifies as an expert in this case. He holds a Ph.D. in human experimental psychology with a sub-specialty in methodology and statistics. He has taught courses in psychology, statistics, and the scientific analysis of data. 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.II p. 377-78; Ex. 359 App. A; Court Order 12/31/03 (Docket 204). The court finds that Dr. Cole’s education, experience, knowledge, and skill qualify him to testify as an expert in this case. 2. Dr. Zax’s Analysis Defendants’ expert, Dr. Zax, used ecological inference (El)- Ex. 948. El estimates turnout and vote shares for two groups within a district and within each precinct of that district. The basis for estimates of turnout and voting behavior for white and minority voters is referred to as “Goodman’s regression,” and El derives estimates from these parameters. Ex. 948 p. 7. El estimates turnout rates and the proportion of voters choosing a particular candidate within the feasible range. The estimates will not be less than zero or greater than one, and the parameters may be much narrower than this range. El calculates these bounds and ensures that the estimated turnout rates and voting shares he within them. Ex. 948 p. 8. Tables two and three in Dr. Zax’s report provide examples of the bounds and estimates, drawn from the analysis of the 2000 General Election in state senate District 27. The first two columns in Table 2 provide lower and upper bounds for the proportion of Indian voters in each precinct choosing the Democratic candidate in the race. Ex. 948 p. 8. When Indians make up only a small proportion of the electorate and prov