Full opinion text
MEMORANDUM AND OPINION ROSENTHAL, District Judge. TABLE OF CONTENTS The Legal Framework..........................................................1200 Findings of Fact................................................................1202 I. The Parties.....................................;......................1202 II. The Background of this Lawsuit..........................................1203 III. The Fact and Expert Witnesses..........................................1204 IV. The First Gingles Factor................................................1204 A. The 1990 Census: Distribution of Hispanic Voting-Age Population in the PISD....................... 1204 B. The Proposed Single-Member District Plans...........................1205 C. Gingles I: A Sufficiently Large and Compact Group of Voting-Age Citizens...........................................................1208 D. Gingles I: 1990 Census Figures on Hispanic Voting-Age Citizens.........1209 E. Gingles I and Projected Population Trends ............................1210 1. Reliability Problems in the 1990 Census...........................1210 2. Reliability Problems in the 1997-98 Projections.....................1211 3. The Least Unreliable Alternative................................1212 V. The Second and Third Gingles Factors....................................1213 A. Analysis of PISD Board Elections ....................................1213 1. The 1987 Election (Position 2)...................................1216 2. The 1987 Election (Position 3)...................................1216 3. The 1988 Election (Position 5)...................................1217 4. The 1989 Election (Position 6)...................................1217 5. The 1990 Election (Position 2)...................................1217 6. The 1993 Election (Position 1) ...................................1218 7. The 1993 Election (Position 2)...................................1218 8. The 1993 Election (Position 3)...................................1218 9. The 1994 Election (Position 4)...................................1219 10. The 1994 Election (Position 5)...................................1219 11. The 1995 Trustee Election......................................1219 12. The 1996 Election (Position 2)...................................1219 13. The 1997 Election (Position 5)...................................1220 B. Gingles II: The Political Cohesion of the Minority Group................1220 1. The Limitations of the Statistical Election Analysis.................1220 2. Hispanic Political Cohesion......................................1221 C. Gingles III: Legally Significant Bloc Voting...........................1221 VI. The Zimmer Factors.................................. 1222 A. Voting Barriers ....................................................1222 B. Candidate Slating Process...........................................1223 C. Vestiges Within the School System....................................1224 D. Socioeconomic Factors..............................................1224 E. Exogenous Elections................................................1225 1. Regression Analyses of Exogenous General Elections...............1225 2. General Analysis of the Exogenous Elections......................1227 F. Racial Appeals.....................................................1227 G. Responsiveness of the PISD .........................................1227 H. The Value of the State’s Interest.....................................1227 VII. Evidence of Invidious Discriminatory Intent.......:........................1228 Conclusions of Law.............................................................1228 I. Gingles I: A Sufficiently Large and Geographically Compact Group...........1228 II. Gingles II: Political Cohesion of the Minority..............................1229 III. Gingles III: Legally Significant Bloc Voting.................. 1229 IV. The Zimmer Factors....................................................1229 V. The Totality of the Circumstances ........................................1280 VI. The Fourteenth and Fifteenth Amendments................................1230 VIL Conclusion Appendix................................ .....................................1231 In this voting rights case, plaintiffs challenge the at-large voting system for electing members of the school board for the Pasadena Independent School District (the “PISD”). Plaintiffs contend that the current system dilutes the votes of politically cohesive Hispanics in the PISD, in violation of section 2 of the Voting Rights Act, 42 U.S.C. § 1973, et seq., and the Fourteenth and Fifteenth Amendments to the United States Constitution. The parties presented evidence to the court from May 31, 1995 through June 8, 1995. Because significant legal developments occurred after that date, this court reopened the record on February 10, 1997 to hear arguments and evidence on elections and relevant demographic changes which had taken place since June 1995. “Because the resolution of a voting dilution claim requires close analysis of unusually complex factual patterns, and because the decision of such a case has the potential for serious interference with state functions, ... district courts [must] explain with particularity their reasoning and the subsidiary factual conclusions underlying their reasoning.” Westwego Citizens for Better Gov’t v. City of Westwego, 872 F.2d 1201, 1203 (5th Cir.1989); quoting Velasquez v. City of Abilene, 725 F.2d 1017, 1020 (5th Cir.1984). In accordance with the requirements of Fed. R. Civ. P. 52(a) and the Fifth Circuit precedents, this court examines the applicable law and sets out its Findings of Fact and Conclusions of Law. THE LEGAL FRAMEWORK The fundamental question in this section 2 vote dilution case is whether, as a result of the at-large system for the PISD school board elections, Hispanics in the PISD “do not have an equal opportunity to participate in the political processes and to elect candidates of their choice.” Thornburg v. Gingles, 478 U.S. 30, 44-45, 106 S.Ct. 2752, 2763, 92 L.Ed.2d 25 (1986). Section 2 of the Voting Rights Act, 42 U.S.C. § 1973, as amended, provides that: (a) No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or col- or.... (b) A violation of subsection (a) of this section is established if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens ... in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. The extent to which members of a protected class have been elected to office in the State or political subdivision is one circumstance which may be considered: Provided, that nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion of the population. 42 U.S.C. § 1973 (emphasis in original). Section 2 claims brought against at-large voting, or “multimember” election schemes, are governed by Thornburg v. Gin gles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). In Gingles, the Supreme Court held that the use of at-large voting would not impede “the ability of minority voters to elect representatives of their choice” unless there is a white bloc voting majority that would “usually be able to defeat candidates supported by a politically cohesive, geographically insular minority group.” 478 U.S. at 47-48, 106 S.Ct. at 2765 (emphasis added). The plaintiffs must meet the Gingles three-part threshold test: [j]irst that the group is sufficiently large and geographically compact to constitute a majority in a single member district; second, that it is politically cohesive and third, that the white majority votes sufficiently as a bloc to enable it usually to defeat the minority’s preferred candidate. Growe v. Emison, 507 U.S. 25, 39-41, 113 S.Ct. 1075, 1084, 122 L.Ed.2d 388 (1993) (emphasis added); see also Gingles, 478 U.S. at 49-53, 106 S.Ct. at 2766-67; Clark v. Calhoun County, Miss., 88 F.3d 1393, 1395 (5th Cir.1996). If the Gingles three-part threshold is met, plaintiffs must then also show that under the “totality of the circumstances,” plaintiffs do not possess the same opportunities to participate in the political process and elect representatives of their choice enjoyed by other voters. Rollins v. Fort Bend Independent School District, 89 F.3d 1205 (5th Cir.1996); Clark, 88 F.3d at 1396 (“it will be only the very unusual case in which the plaintiffs can establish the existence of the three Gingles factors but still have faded to establish a violation of § 2 under the totality of the circumstances”); LULAC v. Clements, 999 F.2d 831, 849 (5th Cir.1993) (en bane), cert. denied, 510 U.S. 1071, 114 S.Ct. 878,127 L.Ed.2d 74 (1994). To make this determination, a court must conduct a “searching practical evaluation” of the factors set forth in Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir.1973) (en banc), aff'd sub. nom., East Carroll Parish School Board v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976), and in the Senate Report of the 1982 Amendments to the Act. The Zimmer factors are: 1. the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process; 2. the extent to which voting in the elections of the state or political subdivision is racially polarized; 3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti-single shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group; 4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process; 5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process; 6. whether political campaigns have been characterized by overt or subtle racial appeals; and 7. the extent to which members of the minority group have been elected to public office in the jurisdiction. S. Rep. No. 417, 97th Cong., 2d. Sess., reprinted in 1982 U.S.C.C.A.N. 177 at 206-07; see also Brewer v. Ham, 876 F.2d 448, 451 n. 4 (5th Cir.1989); Westwego, 872 F.2d at 1204-05. Zimmer and the Senate Report recognize two additional factors that may have probative value in determining whether there is a violation of the Voting Rights Act: (1) whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group; and (2) whether the policy underlying the state or political subdivision’s use of such voting qualification, prerequisite to voting, or standard, practice, or procedure is tenu- • ous. Brewer, 876 F.2d at 451 n. 4; Westwego, 872 F.2d at 1204-05. The plaintiffs must show that “under the totality of circumstances, they do not possess the same opportunities to participate in the political process and elect representatives of their choice enjoyed by other voters.” Clark, 88 F.3d at 1395; Westwego, 872 F.2d at 1206. The “totality” test requires a careful evaluation of the Gingles as well as the Zimmer factors. Monroe v. City of Woodville, Miss., 881 F.2d 1327, 1334 (5th Cir. 1989), cert. denied, 498 U.S. 822, 111 S.Ct. 71, 112 L.Ed.2d 45 (1990). In this evaluation, a court must recognize two themes that have been emphasized in similar cases: (1) courts may not tolerate political systems that effectively exclude minority voters from the democratic process; and (2) courts have rejected the notion that the Act secures to any group of citizens the right to obtain political representation in proportion to its numbers. Williams v. City of Dallas, 734 F.Supp. 1317, 1320 (N.D.Tex.1990). The Fifth Circuit has made it clear that requiring race-based remedies under section 2 of the Voting Rights Act is consistent with the Supreme Court’s decisions in Miller v. Johnson, — U.S.-, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995), Shaw v. Hunt, — U.S. -, 116 S.Ct. 1894, 135 L.Ed.2d 207 (1996), and Bush v. Vera, — U.S.-, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996). Clark, 88 F.3d at 1402-08. In Clark, the Fifth Circuit stated that the three Supreme Court decisions “establish a number of important propositions. First, race-based redistricting, even that done for remedial purposes, is subject to strict scrutiny. Second, compliance with section 2 of the Voting Rights Act is a compelling governmental interest. Third, the State must have a strong basis for concluding that the three Gingles preconditions exist in order to claim that its redistricting plan is reasonably necessary to comply with section 2. Fourth, a tailored response to a found violation must use race at the expense of traditional political concerns no more than is reasonably necessary to remedy the wrong.” Id. at 1405-06. Section 2 of the Voting Rights Act is consistent with the Equal Protection Clause of the Fourteenth Amendment even though, by definition, the statute uses race as a basis for remedial districts. Id. at 1408. “Redistricting to remedy found violations of section 2 of the Voting Rights Act by definition employs race. Miller, Shaw, and Bush, ... do not foreclose the ability of States to remedy the reality of racial inequality in our political system____ The limit is that the remedy must use race at the expense of traditional political concerns no more than is reasonably necessary to remedy the found wrong.” Id. at 1408 (internal quotations omitted). In order to establish a cause of action under the Fourteenth and Fifteenth Amendments, plaintiffs must establish that the defendants acted with a discriminatory design or purpose. Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 264-66, 97 S.Ct. 555, 563, 50 L.Ed.2d 450 (1977); Washington v. Davis, 426 U.S. 229, 238-42, 96 S.Ct. 2040, 2047-48, 48 L.Ed.2d 597 (1976). FINDINGS OF FACT I. The Parties The individual plaintiffs — Rosalinda Perez, Dolores E. Garcia, Maggie Ramirez, Zina Gonzales, Maria Gonzales, Celestino M. Perez, Jr., Jenky M. Diaz, David R. Segura, Rudy N. Trevino, and Yvonne Ruth — are Hispanie/Mexican-American citizens of the United States who are residents of the PISD. Plaintiff Celestino Perez unsuccessfully ran for election to the Pasadena City Council, District G in 1993, Position No. 4 on the PISD Board of Trustees in 1994, and the Board of the San Jacinto Junior College District in 1995. Plaintiff, Pasadena Citizens for Equitable Representation, is an informal association comprised of the individual plaintiffs. The association lacks officers, by-laws, membership criteria, and formal organization. The PISD is a political subdivision of the State of Texas, organized in 1889. The PISD covers approximately eighty-five square miles, located in the southwest part of Harris County, Texas, immediately southeast of Houston, Harris County, Texas. The PISD encompasses portions of the incorporated cities of Pasadena, South Houston, Houston, and unincorporated areas of Harris County, Texas. During the 1991-1992 school year, the PISD had approximately 38,671 students attending 51 schools and had approximately 4,000 employees. From 1981 to 1994, the PISD experienced growth at an annual rate of approximately five to six percent in its average daily attendance population. According to the Texas Education Agency 1991-1992 fall survey, of the approximately 38,671 students enrolled in PISD, forty-nine percent (49%) were White/Caucasian, forty-two percent (42%) were Hispanie/Mexiean-Ameriean, five percent (5%) were African-American, and four percent (4%) were “other” non-white. The Pasadena School Board of Trustees (the “Board”) is responsible for the management and governance of the PISD under Tex. Educ. Code Ann. §§ 23.01-23.999 (Vernon 1987 & Supp.1996). The PISD Board of Trustees is comprised of seven members, each elected at-large by voters living in the PISD. In 1992, the individual defendants— Carmen Orozco, Denny Delafield, Vickie Morgan, Bob Blair, Marshall Kendricks, Harvey Turner, and John Elam — were the elected members of the Board of Trustees of the PISD. Defendant Carmen Orozco, a Hispanic/Mexican-American, first won election to the PISD Board of Trustees in 1987. Orozco was reelected in 1990, 1993, and 1996. Orozco ran unopposed in 1996. II. The Background of this Lawsuit Members of the PISD Board of Trustees are elected through an at-large system. Each of the seven PISD Trustees is elected for a staggered three-year term. Two or three positions are filled by district-wide elections each year. Each Trustee is elected by a plurality of the votes cast on a district-wide basis. Candidates run for a specific position on the Board. Each qualified voter in the PISD is entitled to vote for one candidate for each position up for election in any given year. The elections for the Board of Trustees of the PISD are non-partisan. In 1986, the PISD Board voluntarily studied converting from the at-large system to a single-member district system of electing Trustees. The PISD Board appointed a committee to study the potential conversion and took comments from the public. One group that provided comments, the Hispanic Advisory Committee, consisted of approximately twenty to thirty members of the Hispanic community and was represented by Isaiah Torres. The PISD Board held public hearings and community meetings regarding the adoption of a single-member district election system. The Board retained a demographer, Bobby Bowers, who drew district election lines for proposed single-member districts. The PISD Board adopted a “5-2 plan” using Bowers’ district lines and submitted the plan to the United States Department of Justice for preclearance, to use by January 1987. The plan did not receive support from the Hispanic community or from the community at large. The Hispanic Advisory Committee opposed the plan on the grounds that the plan divided the Hispanic community and did not take into account the Hispanic growth rate. The Department of Justice declined to preelear the 1986 proposed 5-2 plan. The PISD Board withdrew the proposed plan and rescinded its order adopting the 5-2 plan. On November 19, 1992, plaintiffs filed this suit, claiming that Hispanies do not have an equal opportunity to participate in the political process and to elect to the PISD Board candidates of then choice, because of the at-large election system; the use of staggered terms and the numbered post system which eliminate single-shot voting; the large population of the district; the comparatively small number of polling places; the absence of minorities as election officials; and the economic differences between the Anglo and minority communities. Plaintiffs ask this court to require the PISD to change the method of electing Trustees from the at-large system to a single-member district system. The PISD Board of Trustees voted unanimously to defend this lawsuit. III. The Fact and Expert Witnesses Plaintiffs offered the testimony of Dr. Henry Flores as an expert in statistical analysis. Dr. Flores holds B.A., M.A., and Ph.D. degrees in political science, and is the chair of the Department of Political Science at St. Mary’s University in San Antonio, Texas. Plaintiffs also offered the testimony of George Korbel as an expert witness on the Gingles threshold, as well as on the Zimmer factors. Korbel holds a B.A. degree in government and philosophy and a J.D. degree from the University of Minnesota. Since 1981, Korbel has worked as a litigation coordinator for Texas Rural Legal Aid, Inc. Korbel is a recognized voting rights expert, whose focus is on the non-statistical analysis of elections and the factors that influence their outcome. Plaintiffs also offered the testimony of John E. Castillo (“Castillo”), who has experience as a political consultant, is familiar with the demographics of South Houston and Pasadena, and has worked on redistricting plans for municipal elections in Houston. Mr. Castillo provided estimates as to the numbers of Hispanics of voting age living in the PISD. At the February 10, 1997 hearing, plaintiffs offered the testimony of Dr. Tatcho Mindiola, Jr. as an expert on the first Gingles threshold. Dr. Mindiola updated the demographic projections offered by Mr. Korbel in the 1995 trial, using Mr. Korbel’s methodology. Dr. Mindiola holds a Bachelor of Business Administration and M.A. and Ph.D. degrees in sociology. Dr. Mindiola is an associate professor of sociology at the University of Houston and director of the University of Houston Center for Mexican American Studies. Defendants offered the testimony of Dr. John Alford as an expert in statistical analysis, to explain the significance of the voting patterns and to examine population growth trends. Dr. Alford holds B.S., M.A., and Ph.D. degrees in political science and an M.P.A. in public administration. He is an associate professor at Rice University. The court finds that Drs. Flores, Mindiola, and Alford are experts in the field of statistical analysis; that Drs. Flores and Alford are experts in applying statistical analysis to issues raised by election districts; and that Mr. Korbel is an expert in the field of non-statistical voting rights analysis. Mr. Castillo is a fact witness regarding elections in the South Houston and Pasadena areas. IV. The First Gingles Factor The threshold question is whether it is possible to draw one or more majority-minority districts in the PISD. Drawing majority-minority districts requires a minority population that is of sufficient size and compactness to allow the construction of a compact and contiguous district in which the minority group can form an effective voting majority. A. The 1990 Census: Distribution of Hispanic Voting-Age Population in the PISD According to the 1990 United States Census of Population, the PISD had 187,047 residents, of whom 62.07 percent are Anglo (116,104); 30.22 percent are Hispanic (56,-529); 3.94 percent are African-American (7,371); and 3.74- 77 percent are “other” (7,003). The voting-age population of the PISD according to the 1990 census is 131,-405, of whom 66.84 percent are Anglo (87,-824); 26.06 percent are Hispanic (34,247); 3.56 percent are African-American (4,671); and 3.56 percent are “other” (4,675). The 1990 census population and voting-age population (“VAP”) figures and percentages are reflected in the following table: PISD POPULATION PISD VAP Number of Residents Percentage of PISD Population Number of Residents Percentage of PISD Population Anglo 116,104 62.07% 87,824 66.84% Hispanic African-56,529 30.22% 34,247 26.06% American 7,371 3.94% 4,671 3.56% “Other” 7,003 3.74-3.77% 4,675 3.56% Total 187,047 100% 131,405 100% There are only four polling places in the PISD for Board elections. The polling places are located at Pasadena High School, Sam Rayburn High School, South Houston High School, and J. Frank Dobie High School. The geographic attendance zones for each of the four schools serve as the boundary lines for each polling place. The population and voting data for the 1986 through 1997 elections is grouped by the four polling places. For the 1987-1994 elections, the areas served by the four polling places show the following distribution of Hispanic voting-age population: HISPANIC PERCENTAGE OF VAP IN POLLING PLACE EACH POLLING PLACE Pasadena High School 35.33% South Houston High School 30.97% Rayburn High School 20.80% Dobie High School 15.36% B. The Proposed Single-Member District Plans In this lawsuit, plaintiffs argue that it is possible to develop a plan for seven single-member districts using 1990 census data (the “Korbel Plan”). Under this plan, in proposed District 1, 55.07 percent of the voting-age population would be Hispanic and in proposed District 6, 50.04 percent of the voting-age population would be Hispanic. The Korbel Plan does not take into account the citizenship of the voting-age Hispanics; projected population growth trends in the Hispanic community; or a possible undereount of Hispanics in the 1990 census. The Korbel Plan was not intended as a specific “remedy district,” but was offered to show the feasibility of developing a single-member district plan meeting the first Gingles requirement. Plaintiffs also presented a proposed plan for five single-member districts and two at-large districts (the “5-2 Plan D”). The 5-2 Plan D did not take into account the citizenship of the voting-age Hispanics; projected population growth trends in the Hispanic community; or a possible undercount of Hispanics in the 1990 census. The 5-2 Plan D contained one district, (District 1), consisting of 51.9 percent total Hispanic population and 46.0 percent Hispanic voting-age population, and another district, (District 2), consisting of 47.0 percent total Hispanic population and 41.9 percent Hispanic voting-age population. Districts 1 and 2 of the 5-2 Plan are similar in location and configuration to the two proposed Hispanie-majority districts in the Korbel Plan. No projections of the populations of the districts in the 5-2 Plan D were presented. In 1994, a PISD-retained demographer, Leslie Johnston (“Johnston”), developed a plan for seven single-member districts (the “Johnston Plan”), also using 1990 census data. The Johnston Plan contained one district, (District 1), consisting of a total Hispanic population of 58.4 percent and a Hispanic voting-age population of 52.5 percent, and a second district, (District 2), made up of a total Hispanic population of 49.2 percent and a Hispanic voting-age population of 44.0 percent. Johnston also developed two plans with seven single-member districts, each showing one district of over 50 percent Hispanic voting-age population. “Plan A” included one district, (District 1), with a Hispanic population of 58.8 percent and a Hispanic voting-age population of 52.9 percent and another district, (District 2), with a Hispanic population of 50.3 percent and a Hispanic voting-age population of 45.2 percent. “Plan B” contained one district, (District 1), consisting of a Hispanic population of 58.6 percent and a Hispanic voting-age population of 52.7 percent and another district, (District 2), consisting of a Hispanic population of 48.9 percent and a Hispanic voting-age population of 43.8 percent. The Johnston Plan did not take into account the citizenship of the Hispanic voting-age population in the district; population growth estimates; or a possible under-count of Hispanics in the 1990 census figures. Between 1980 and 1990, the total Hispanic population in the PISD increased by 80.55 percent (25,131); the total African-American population increased by 90.79 percent (3,146); and the total Anglo population decreased by 16.16 percent (-22,212). This represents an eight percent annualized increase in the total Hispanic population and a one to two percent annualized decrease in the Anglo population of the PISD from 1980 to 1990. At each of the four high schools within PISD, the percentage of Hispanic students grew and the percentage of Anglo students dropped from 1981 to 1994. Plaintiffs presented evidence that the application of simple annualized growth rates of 8 percent for Hispanics, 9 percent for African-Americans, and negative 1 to 1.5 percent for Anglos, results in the following voting-age population (“VAP”) projection estimates for the proposed single-member districts: Korbel Plan HISPANIC POPULATION MEASURED DISTRICT 1 DISTRICT 6 1990 VAP 55.07% 50.04% 1990 CITIZEN VAP Not Available Not Available 1995 VAP* 64.63% 58.50% 1995 CITIZEN VAP* 55.92% 47.48% 1996 VAP* 66.21% 59.90% 1996 CITIZEN VAP* 57.11% 48.96% 1997 VAP* 67.68% 61.22% 1997 CITIZEN VAP* 58.40% .50.21% 1998 VAP* 69.08% 62.47% 1998 CITIZEN VAP* 60.00% 51.57% ' denotes projections. Johnston Plan HISPANIC POPULATION MEASURED DISTRICT 1 DISTRICT 2 1990 VAP 52.48% 44.01% 1990 CITIZEN VAP Not Available Not Available 1995 VAP* 60.85% 38.90% 1995 CITIZEN VAP* 52.89% 29.80% 1996 VAP* • 62.56% 41.14% 1996 CITIZEN VAP* 54.6 31.08% * denotes projections. Plan HISPANIC POPULATION MEASURED DISTRICT 1 DISTRICT 2 1990 VAP 52.94% 46.38% 1990 CITIZEN VAP Not Available Not Available 1995 VAP* 62.67% 54.43% 1995 CITIZEN VAP* 54.81% 44.33% 1996 VAP* 64.27% 55.73% 1996 CITIZEN VAP* 56.50% 45.63% * denotes projections. Plan B HISPANIC POPULATION MEASURED DISTRICT 1 DISTRICT 2 1990 VAP 52.74% 43.77% 1990 CITIZEN VAP Not Available Not Available 1995 VAP* 62.49% 51.91% 1995 CITIZEN VAP* 54.61% 41.84% 1996 VAP* 64.09% 53.24% 1996 CITIZEN VAP* 56.31% 43.15% : denotes projections. C. Gingles I: A Sufficiently Large and Compact Group of Voting-Age Citizens Plaintiffs argue that to satisfy the first of the Gingles threshold requirements, they need only establish that it is possible to create one single-member district in which the majority of the voting-age population is Hispanic. Defendants contend that plaintiffs must establish that the Hispanic community in the PISD is sufficiently large and geographically compact to create a single-member district in which the majority of voting-age citizens would be Hispanic. The majority of courts addressing this issue have held that a plaintiff must establish that it is possible to create a district with a majority of minority voting-age citizens. McNeil v. Springfield Park District, 851 F.2d 937, 944-45 (7th Cir.1988), cert. denied, 490 U.S. 1031, 109 S.Ct. 1769, 104 L.Ed.2d 204 (1989); Dickinson v. Indiana State Election Board, 933 F.2d 497, 503 (7th Cir.1991) (“a simple majority of eligible voters in the single-member district”); Romero v. City of Pomona, 665 F.Supp. 853, 854 (C.D.Cal.1987), aff'd, 883 F.2d 1418, 1425-26 (9th Cir.1989) (stating that Gingles refers to effective voting majorities, rather than raw population totals, as the touchstone for determining geographical compactness); Morris v. City of Houston, 894 F.Supp. 1062, 1065 (S.D.Tex.1995) (“using data of voting age Hispanic citizens is the correct measure of the Hispanic population’s ability to create a majority voting district”) (appeal pending, No. 95-20740, oral argument heard on 9/30/96); Milwaukee Branch of the NAACP v. Thompson, 929 F.Supp. 1150 (E.D.Wis. 1996) (“citizenship rates can be taken into account in resolving the issue whether a plaintiff has satisfied the first Gingles condition[; individuals who are not citizens are not eligible voters[; tjhose ineligible to vote are not among those whose votes have been diluted within the meaning of section 2”); Skorepa v. City of Chula Vista, 723 F.Supp. 1384,1388-89 (S.D.Cal.1989); contra LULAC v. North East ISD, 903 F.Supp. 1071, 1084 (W.D.Tex.1995). Gingles I requires a showing of a group of voting-age citizens sufficiently large and geographically compact to constitute a majority in the single-member district. This requirement is in harmony with the language of section 2 of the Voting Rights Act. The Act forbids voting qualifications, prerequisites, standards, practices, and procedures which “are not open to participation by members of a class of citizens ... in that its members have less opportunity than other members of the electorate to participate in the political process,” and which result “in a denial or abridgement of the right of any citizen of the United States.” 42 U.S.C. § 1973 (emphasis added). The Supreme Court in Gingles emphasized that “[ujnless minority voters possess the potential to elect representatives in the absence of the challenged structure or practice, they cannot claim to have been injured by that structure or practice.” Gingles, 478 U.S. at 51 n. 17, 106 S.Ct. at 2767 n. 17 (emphasis added); Campos v. City of Baytown, Tex., 840 F.2d 1240, 1244 (5th Cir. 1988), cert. denied, 492 U.S. 905, 109 S.Ct. 3213, 106 L.Ed.2d 564 (1989). Because non-citizens are not eligible to vote, they “have not experienced a dilution of their vote.” McNeil, 851 F.2d at 944-45. Gingles requires a showing based on individuals of voting age to assure that those individuals are voting-eligible. To require citizenship does not add to the requirements already imposed by Gingles. This court concludes that plaintiffs must demonstrate that it is feasible to draw a geographically compact district in which Hispanics comprise the majority of voting-age citizens. D. Gingles I: 1990 Census Figures on Hispanic Voting-Age Citizens The census does not report the number of citizens by race or ethnicity at the block or tract level, as it does for voting-age population. Hispanic voting-age citizenship rates, however, can be estimated by using regression analysis to compare the Hispanic population with citizenship rates. Regression is a mathematical technique used for estimating the best fitting straight line that could be drawn to describe the relationship between two variables in a scatter plot. In a two-variable regression, R is the same as a Pearson’s r or correlation coefficient. R indicates the degree of linear association between the two variables examined. The R range is 0 to 1.0, positive or negative. R2 describes how tightly the actual data points cluster around the regression line. It is the proportion of variation in the dependent variable that is explained or accounted for by the independent variable. The R2 ranges from 0 to 1.0, but is always positive. The application of the regression analysis to the 1990 census reveals that approximately 60 percent of the voting-age Hispanics in the PISD are citizens. An independent estimate, based on survey data gathered by a national Hispanic political organization, reported a 54 percent adult Hispanic citizenship rate for the Houston area. Assuming a 60 percent citizenship rate for voting-age Hispanics in the PISD, a district must have a Hispanic voting-age population of at least 62.5 percent to produce a district with a 50 percent Hispanic citizen voting-age population. Using the 1990 census data, none of the districts in the proposed plans has a Hispanic voting-age population of at least 62.5 percent, or a citizen voting-age population of at least 50 percent. The Korbel Plan District 1 has the highest 1990 Hispanic voting-age population, at 55.07 percent. The parties appear to concede that the 1990 Hispanic citizen voting-age population is not sufficient for Gingles I. The Korbel Plan proposed districts are irregularly shaped, evidencing a lack of geographic compactness. Dr. Alford testified that the Korbel Plan districts were irregular in shape; there were “fingers” extending out of the Korbel Plan districts, with many intrusions only a block wide in size. Dr. Alford testified that this did not appear justified as an attempt to recognize neighborhood boundaries or existing precincts, and was instead an effort to include as many Hispanics as possible given the dispersed residential patterns in the PISD. Korbel did not refute the irregular shape of the proposed districts, but explained that they were not intended as remedy districts. Even if the proposed single-member districts are not “remedy” districts, they must demonstrate the feasibility of drawing single-member districts with similar population sizes and compact shapes. To determine compactness, one “creates hypothetical ideal districts (ie., districts that contain precisely the same number of people) and then determines, in percentage terms, the degree of deviation between each of the actual districts and the ideal one. The maximum deviation is calculated by adding the percentage points that the largest district is above the ideal, to the percentage points the smallest is below.” Garza v. County of Los Angeles, 918 F.2d 763, 785 (9th Cir.1990) (Kozinski, J., concurring in part, dissenting in part), cert. denied, 498 U.S. 1028, 111 S.Ct. 681, 112 L.Ed.2d 673 (1991); citing Brown v. Thomson, 462 U.S. 835, 838-39, 103 S.Ct. 2690, 2694, 77 L.Ed.2d 214 (1983). In order not to violate the one man one vote principle, the maximum deviation from the smallest to the largest district should not exceed 10 percent. Dr. Alford testified that the Korbel Plan had a population deviation of approximately 32 percent. Korbel did not refute this testimony. The proposed districts are flawed by a citizenship population of less than 50 percent, a population deviation of over 10 percent, and irregular shapes. Using the 1990 census data, plaintiffs fail to prove that it is feasible to create one or more districts in which a majority of voting-age citizens are Hispanic. E. Gingles I and Projected Population Trends Because plaintiffs cannot satisfy Gingles I using the 1990 census data, they ask this eourt to determine the feasibility of creating a majority-minority district using projected estimates for the 1995 through 1998 Hispanic voting-age and citizen voting-age population in the PISD. A court is not confined to United States census data in deciding whether a sufficiently large and geographically compact majority-minority single-member district can be created to satisfy Gingles I. Kirkpatrick v. Preisler, 394 U.S. 526, 534-36, 89 S.Ct. 1225, 1231, 22 L.Ed.2d 519 (1969); Garza, 918 F.2d at 772-73; Dixon v. Hassler, 412 F.Supp. 1036, 1040 (W.D.Tenn.), aff'd, 429 U.S. 934, 97 S.Ct. 346, 50 L.Ed.2d 303 (1976); see also Westwego Citizens for Better Gov’t v. City of Westwego, 906 F.2d 1042, 1045-46 n. 3 (5th Cir.1990). However, the census figures are presumed accurate until proven otherwise. McNeil, 851 F.2d at 946. Proof of changed figures must be “thoroughly documented,” have a high degree of accuracy, and be “clear, cogent and convincing” to override the presumptive correctness of the prior decennial census. McNeil, 851 F.2d at 946; Kirkpatrick, 394 U.S. at 534-36, 89 S.Ct. at 1231; Dixon, 412 F.Supp. at 1040; Graves v. Barnes, 446 F.Supp. 560, 568 (W.D.Tex. 1977), aff'd, 435 U.S. 901, 98 S.Ct. 1444, 55 L.Ed.2d 492 (1978); see also Dickinson, 933 F.2d at 503. Federal census data must be used unless there is more reliable data. McNeil, 851 F.2d at 946; Skorepa, 723 F.Supp. at 1390. “Estimates based on past trends are generally not sufficient to override ‘hard’ decennial census data.” McNeil, 851 F.2d at 946. 1. Reliability Problems in the 1990 Census Korbel testified that minorities were undercounted in the 1990 census by approximately ten percent in Harris County. The concern over the undereount has been reviewed by federal courts. The issue is whether the asserted undereount of Hispanic voters is supported by “concrete evidence” sufficient to override the presumption of accuracy given the census data. Skorepa, 723 F.Supp. at 1390; Cuomo v. Baldrige, 674 F.Supp. 1089, 1104-05 (S.D.N.Y.1987). In Wisconsin v. City of New York, — U.S. -, 116 S.Ct. 1091, 134 L.Ed.2d 167 (1996), the Supreme Court discussed at length the concern over “a net undereount of the actual American population in every decennial census,” stated that the 1990 census was “successful in counting 98.4% of the population,” and held that it was reasonable for the Secretary of Commerce to conclude that the “actual enumeration” mandated by the Constitution could best be achieved in the 1990 census without the postenumeration survey statistical adjustment designed to correct an undereount in the initial enumeration. This court finds the alleged under-count to be an insufficient basis for rejecting the 1990 census as unreliable and for requiring estimates projecting population growth in the proposed districts. Plaintiffs offered no evidence as to any specific effect of the alleged undereount on the size of the Hispanic adult citizen population, or on the non-Hispanie population, in any of the proposed or actual districts. Dr. Alford noted that both Hispanics and non-Hispanies could have been undercounted, and that noncitizen Hispanics could well constitute the majority of those not counted. Plaintiffs have not cited any case in which a court has rejected the census figures in a voting rights case based on an alleged undereount. To the contrary, several courts have rejected such a claim for lack of evidence that the undereount made the original census count insufficiently reliable. See, e.g., Skorepa, 723 F.Supp. at 1390; Cuomo, 674 F.Supp. at 1104-05. Common sense, as well as the expert testimony, teaches that the further away in time one moves from the census, the less reliable it becomes. However, Drs. Alford and Minchóla agreed that the further away one moves from the census, population projections based on the census also become more unreliable. The parties, and their experts, agree that the 1990 census count of the Hispanic voting-age population in the relevant area is outdated. The experts agree that the population has grown. The experts disagree as to whether the estimated projections of that growth in the proposed single-member districts in the PISD are more reliable than the 1990 census count. 2. Reliability Problems in the 1997-98 Projections The plaintiffs’ experts used a decenniallycompounded “straight-line” estimate to make their population projections. Plaintiffs’ experts calculated the total growth rate between 1980 and 1990, then annualized it and used a straight-line growth rate between 1980 and 1990. In 1990, plaintiffs changed the population base used from the 1980 population to the 1990 population, adding the 1980-1990 population growth to the new 1990 base. Between 1990 and 2000, plaintiffs again used a straight-line growth rate. Dr. Alford testified that plaintiffs’ linear progression was not a single straight-line estimate between 1980 and 2000, but two straight lines connected in 1990. The 1990-2000 straight line has a steeper slope than the 1980-1990 straight line. Dr. Alford testified that the decennially compounded straight-line estimate for population growth, which plaintiffs’ experts used to calculate the projected Hispanic population in the proposed districts, is overly simplified and wholly unreliable as a measure of the 1997 Hispanic voting-age citizenship population. Other factors must also be taken into account, such as age cohorts, birth and death rates, in- and out-migration, and absolute limiting factors, such as housing availability in a particular geographic area. Dr. Mindiola agreed that linear progressions had shortcomings in terms of accuracy, but were simple, easy to calculate, and more practical than extensive surveys. Courts have recognized that demography is complex. Courts have rejected overly simplistic, “crude” analyses that are easy and inexpensive to calculate but too inaccurate to serve as a basis for changing the basis of conducting elections. Dixon, 412 F.Supp. at 1041; see also Kirkpatrick, 394 U.S. at 534-36, 89 S.Ct. at 1231; see also McNeil, 851 F.2d at 946. In Dixon, the court noted that other factors must be taken into account to provide reliable projections, such as population density and land area; current information on the area with respect to listed telephones and registered automobiles; and an area’s housing type, housing value, income and age of people, and family structure. 412 F.Supp. at 1041. Dr. Mindiola testified that the straight-line method used by plaintiffs results in “conservative” population projections, for voting-age populations as well as for citizen voting-age populations. Dr. Mindiola stated that the Hispanic population was growing at a faster rate than the straight-line projection reflected. However, Dr. Mindiola could not identify for this court the basis for this estimate of the Hispanic population growth rate, citing the need for extensive surveys. Dr. Mindiola assumed a high growth rate for the entire Hispanic population in the PISD, but could not identify a growth rate for Hispanic citizens in the proposed districts. Dr. Mindiola agreed with Dr. Alford that the citizenship rate applied to his projections would be lower if the non-citizen growth rate were greater than the citizen growth rate. Dr. Alford testified that the Hispanic citizenship rate was not constant but rather decreased during the 1990s, in part due to an increase in illegal immigration. Dr. Alford testified that simply applying the 1980-1990 growth rate to project the 1996, 1997, and 1998 populations, as plaintiffs did, leads to unreliable results. Dr. Alford supported his concern with specific evidence in the record. If the plaintiffs’ population growth projections were accurate, the percentage of Hispanics voting in the PISD and exogenous elections would have increased; instead, the percentage remained essentially unchanged. If the plaintiffs’ projections were accurate, Democrats, consistently favored by the Hispanic voters in the PISD, would have shown an increase in votes received over tune; instead, the number of votes cast for Democratic candidates by voters in the PISD was stable. Dr. Alford testified that Hispanic voter registration did increase during the 1990s, but not at the rate that would be expected if plaintiffs’ Hispanic citizen growth projections were accurate. Dr. Alford testified that the increase in Hispanic voter registration is consistent with the presence of growth in the Hispanic voting-age citizen population since the 1990 census. However, the likely rate of that growth does not approach the rate plaintiffs project. Dr. Alford testified that population projections are generally highly sensitive to assumptions and prone to error, and not simply “straight line.” Dr. Alford testified that the 1990 census was the best estimate of the PISD populations, despite the passage of time. This court finds that there is no credible evidence that plaintiffs’ straight-line projections for growth in the Hispanic citizen voting-age population in the PISD are sufficiently rehable for the purpose of the Gingles analysis. 3. The Least Unreliable Alternative This court is faced with a Hobson’s choice between two unsatisfactory alternatives. The midpoint on the census cycle is passed. The 1990 census is recognized as inaccurate to a certain extent. However, the only estimates of the extent to which it is unreliable, and the only estimates of the current population, are estimates that are also unreliable. The choice is between an outdated actual count and an estimated projection based in part on that count and in part on assumptions and arithmetic. In McNeil, the Seventh Circuit addressed a similar situation. 851 F.2d 937. In that ease, decided in July 1988, the 1980 census “indicated a minority voting age population of roughly 43 %.” Id. at 945. The plaintiffs-appellants “reified] on their expert’s demographic analysis [to] estimate] a minority population of 56% in the proposed district as of 1987.” Id. The appellants quoted Gaffney v. Cummings, 412 U.S. 735, 745-47, 93 S.Ct. 2321, 2328, 37 L.Ed.2d 298 (1973), in which the Supreme Court stated in dicta that: it must be recognized that total population, even if absolutely accurate as to each district when counted, is nevertheless not a talismanic measure of the weight of a person’s vote under a later-adopted reapportionment plan. The United States census is more of an event than a process. It measures population at only a single instant in time. District populations are constantly changing, often at different rates in either direction, up or down. Substantial differentials in population growth rates are striking and well-known phenomena. The McNeil court agreed with the appellants that, “[w]ith a mobile population, the demographic facts may always be a bit stale.” McNeil, 851 F.2d at 945. However, the court found that the appellants’ argument and the Gaffney dicta suggested only that “census data may not accurately reflect present population.” Id. at 946. The argument did not “require courts to use a different talisman.” Id. The court stated that it would depart from the census and its presumption of accuracy only if the “proof of changed figures” was “clear and convincing.” Id. The court also stated that “[i]n the present state of the demographic art, estimates based on past trends are generally not sufficient to override ‘hard’ decennial census data____ [Populations change constantly and kinks in the data are common.” Id. The court ultimately held that it would use the census figure of 46 percent for the minority voting age population, but added that “if the plaintiffs’ view of the population trends is correct, they will have a good claim based on the 1990 census.” Id. This court finds Dr. Alford’s testimony and criticisms of plaintiffs’ population projections to be credible. Although plaintiffs’ projections may have been cost-effective and simple to calculate, they are unreliable estimates that offer only “a talismanic measure” of the population. See McNeil, 851 F.2d at 945-46 (“in July 1988, the court relied on the 1980 census because it found that ‘the meager [projection] evidence’ before [it could not] override the presumption of census accuracy”); see also Graves, 446 F.Supp. at 568 (in October 1977, the court relied on the 1970 census rather than the plaintiffs’ projections because it found “several scientific defects” in plaintiffs’ method of projection); see also Skorepa, 723 F.Supp. at 1390 (in October 1989, the court used the 1980 census data because the projection estimates offered by plaintiffs were “not based on the kind of concrete evidence necessary to override the presumption of accuracy of the census data”). This court does not find the projected population growth estimated by plaintiffs to be sufficiently reliable to overcome the presumption of census accuracy. If the plaintiffs’ estimate of the population trends is correct, like the plaintiffs-appellants in McNeil, the plaintiffs in this case may have a stronger argument based on the 2000 census. However, using the 1990 census, the most reliable estimate available to this court, plaintiffs fail to meet Gingles I. V. The Second and Third Gingles Factors A. Analysis of PISD Board Elections From 1987 to 1997, there have been twelve contested races for the Board of Trustees of the PISD. Three of those elections involved head-to-head contests. Two of those three head-to-head contests involved a minority candidate. The thrust of plaintiffs’ claim is that white bloc voting defeats the preferred candidate of the minority community unless there are “special circumstances” present. A breakdown of the results of the PISD elections with Hispanic candidates between 1987 and 1997 is summarized below: YEAS (POS.) 1987(2) 1987(3) 1988(5) 1989(6) 1990(2) 1993(1) 1993(2) 1993(3) 1994(4) 1994(5) 1997(5) CANDIDATE (* incumbent) Orozco Stokes Blumrick Total Votes Cast: Elam Gamer * Valdez Total Votes Cast: Blair Robinson Early Total Votes Cast: Kendrick Garvey-Montanez Kowis Rudd Total Votes Cast: Orozco Jackson Total Votes Cast: Roberts Graham Briscoe Gutierrez Total Votes Cast: Orozco * Maxwell Total Votes Cast: N. Sullivan Andrews Leatherman Hell Garcia Crawford Total Votes Cast: T. Sullivan Jordan Castillo Filberth Total Votes Cast: Blair* McClain Bennett Briscoe Segura Total Votes Cast: Blair* Phillips Gonzales Total Votes Cast* R A C E H W W W w w H W W H W W H W W H W W W W H H W W W W W H W W W H H W W W W W W H W W H VOTES 2,428 1,164 1,137 654 5.383 3,293 1,694 386 5,373 1,840 952 522 278 3,592 2,077 196 49 37 2,359 643 149 792 3,393 1,508 1,249 203 6,353 4.383 2,038 6,421 3,984 1,794 246 168 118 54 6,310 2,105 994 247 149 65 47 3,607 1,998 584 534 264 214 3,594 1,243 250 129 1,731 TOTAL 45.10% 100% 61.29% 7.18% 100% 51.22% 14.53% 100% 88.05% 8.31% 100% 81.9% 100% 53.41% 3.20% 100% PERCENTAGE OF VOTES RECEIVED IN: PASADENA S.HOUSTON RAYBURN 45.93% 49.33% 57.22% 100% 63.14% 1.87% 100% 10.98% together 100% 55.59% 5.95% 100% 71.81% 7.45% 4.87% 100% 66.04% 10.63% 100% 90.51% 5.53% 100% 78.61% 100% 49.53% 3.68% 100% 100% 65.65% 100% 54.79% 13.74% together 100% 57.60% 10.40% 100% 84.69% 6.63% 100% 100% 57.89% 100% 35.94% 47.03% 100% 83.72% 13.95% 100% 90.23% 100% 45.39% 4.16% 100% 61.04% 100% 53.59% 2.59% 100% 60.54% 11.07% together 100% 56.73% 5.57% 100% 75.74% 10.66% 100% 73.21% 7.12% 100% 71.07% 9.95% 100% 92.30% 4.13% 100% 74.75% 100% 48.45% 2.90% 100% 68.23% 62.94% 1.09% 100% 57.12% 9.38% together 100% 61.69% 100% 82.67% 6.10% 100% DOBIE 21.79% 100% 35.85% 9.28% 100% 19.09% 4.66% 100% 69.03% 23.51% 100% 85.42% 100% 61.71% 2.18% 100% 66.26% 100% 61.50% 2.26% 100% 50.39% 34.64% together 100% 31.28% 7.05% 100% 64.67% 10.67% 100% Both sides used regression analysis to determine whether there is a pattern of voting preferences in the PISD elections. Rollins, 89 F.3d at 1210 (the Supreme Court and the Fifth Circuit have approved the use of statistical analysis to evaluate racially polarized voting). Drs. Flores and Alford often agreed on the results of applying the statistical techniques to the elections. They disagreed as to the legal significance of the regression analysis results they produced. The regression analyses are one factor that this court examines in determining whether the second and third prongs of the Gingles test are met and whether the Zimmer factors evidence a section 2 violation. They are, however, only one factor among the “totality of the circumstances” that this court must examine. Regression analyses in a voting rights case compare the density of a particular population group in a specified area with the percentage of votes received by a particular candidate in that area. See Houston v. Lafayette County, 56 F.3d 606, 609 n. 1 (5th Cir.1995); see also Teague v. Attala County, 92 F.3d 283, 285 (5th Cir.1996) (“[ecological regression analysis explains how voting and turnout correspond to the proportions of whites and [minorities] in each voting precinct”). In this ease, the regression analysis examines to what extent the vote for a Hispanic candidate increases as the concentration of Hispanic voters in the precinct increases. The coefficients for the y-intercept and for the slope form the equation of the actual regression line. The y-intercept defines the point at which the regression line crosses the vertical axis, giving the predicted value of the votes cast for the Hispanic candidate (the dependent variable) when the percentage of the Hispanic voting-age population (the independent variable) is zero. The slope indicates the rise of the regression line, or the predicted change in the votes cast for_ the Hispanic candidate (the dependent variable) for a one-percentage-point change in the percentage of Hispanic voting-age population (the independent variable). Using both the y-intercept and the slope, the predicted share of the Hispanic candidate’s votes can be estimated in hypothetical “extreme” cases in which the Hispanic voting-age population is 100 percent or 0 percent. See Teague, 92 F.3d at 285 (“[e]xtreme case analysis looks separately at the actual votes [minority] candidates received in the most heavily [minority] and white precincts to give a clear indication of the preference of voters for either [minority] or white candidates in the two types of districts”); see also Rollins, 89 F.3d at 1210 n. 5. Extreme case analyses are generally based on precincts in which one race constitutes 90 to 100 percent of the population. As the percentage of the population analyzed falls below 90 percent, the analysis becomes less reliable and less useful. In this case, none of the four polling areas is one in which one race constitutes 90 to 100 percent of the population. The extreme case analysis conducted for the PISD elections is based on hypothetical extreme precincts and used only as an analytical tool to estimate the support received by each PISD trustee candidate from Anglo and Hispanic voters. The results of Dr. Flores’s regression analysis for the PISD elections are summarized below: Regression Analysis of 1987-1997 PISD Election Results YEAR CANDIDATE R SLOPE Y-INTERCEPT R2 1987 Valdez .598 .208 5.639 .351 Orozco -.0176 -.048 43.989 .0003 1989 Garvey-Montanez .712 • 2.11 -1.550 .512 1990 Orozco .782 .504 76.026 .611 1993 Gutierrez .158 .046 2.901 .025 Orozco -.727 -.858 72.340 .519 Garcia .990 .208 .405 .967 1994 Perez .955 .427 .003 .912 Castillo -.167 -.085 .082 .028 Segura -.726 -8.879 1.050 .527 1995 Hernandez .745 .562 2.421 .555 Perez .826 1.143 1.970 .682 1997 Gonzalez .914 N/A .021 N/A Dr. Flores testified that in his opinion, an R = .4 or above indicates a “good check” of the importance of race in voting behavior. Dr. Flores cited other researchers, Delbert Miller and Earl Babbles, who also consider that an R = .4 indicates a strong relationship. Dr. Alford testified that ordinarily, when .4= <R < = .6, there is a strong correlation between variables, but with smaller data sets this “rule of thumb” does not apply because any occurrence could be the result of random chance. The experts’ analyses, as well as the testimony of the fact witnesses, have been examined in detail by this court. The court’s findings as to the expert and fact witnesses’ testimony are set out below — first, in connection with the analysis of the individual elections, then in connection with how this analysis shapes the Gingles and Zimmer inquiries. 1. The 1987 Election (Position 2) The candidates for Position 2 in the 1987 election were Stokes, Blumriek, Orozco (Hispanic), and Seiler. Orozco was elected, receiving 45.10 percent (2428/5383) of the total vote; 45.93 percent (435/947) of the vote in Pasadena; 49.33 percent (442/896) of the vote in South Houston; 57.22 percent (1058/1849) of the vote in Rayburn; 21.79 percent (277/1271) of the vote in Dobie; and 51.43 percent (216/420) of the absentee vote. Dr. Flores testified that the R for Hispanics voting for Orozco in the 1987 election for Position 2 was -.0176 with a y-intercept of 43.989. The Anglo candidates received a total of 2955 votes. Stokes received 1164 votes; Blumriek received 654; and Seiler received 1137. The Anglo candidates split the white vote. Orozco won the 1987 election for Position 2 by a plurality. Looking at yintereepts and hypothesizing about homogeneous precincts, Dr. Flores estimated that in a hypothetical 100 percent Hispanic district, Orozco would have received 39.1 percent of the vote, while in a 0 percent Hispanic precinct, Orozco would receive 43.9 percent of the vote. Dr. Flores testified that there was virtually no correlation between the race of the voters in any of the four polling places and the votes received by Orozco. Orozco was the preferred minority candidate in 1987 (and in 1990); she was not defeated by bloc voting. 2. The 1987 Election (Position 3) The candidates for Position 3 in the 1987 election were Elam, Valdez (Hispanic), and Garner. Elam was elected, receiving 61.29 percent (3293/5373) of the total vote; 68.96 percent (651/944) of the vote in Pasadena; 57.89 percent (510/881) of the vote in South Houston; 73.21 percent (1347/1840) of the vote in Rayburn; 35.85 percent (460/1283) of the vote in Dobie; and 76.47 percent (325/425) percent of the absentee vote. Valdez received 7.18 percent of the total vote; 4.87 percent of the vote in Pasadena; 8.63 percent of the vote in South Houston; 7.12 percent of the vote in Rayburn; 9.28 percent of the vote in Dobie; and 3.29 percent of the absentee vote. Dr. Flores testified that the R for Hispanics voting for Valdez in the 1987 election for Position 3 was .598. Dr. Flores testified that the relationship betwee