Celine Buckanaga, Francis Gill, Harvey Dumarce, and Nancy Smith, United States citizens of American Indian descent, individually and on behalf of all others similarly situated, appeal from a final judgment entered in the District Court for the District of South Dakota in favor of Sisseton Independent School District No. 54-5 (District), members of the school board, the business manager and superintendent of the District, and the county auditor in their individual and official capacities. The Sisseton-Wahpeton Sioux Tribe has filed an amicus curiae brief. Appellants, registered voters of the District, alleged that the at-large voting system for electing District board members violates § 2 of the Voting Rights Act of 1965, as amended, 42 U.S.C. § 1973. The district court found no violation. Buckanaga v. Sisseton Independent School District, No. 84-1025, slip op. at 9 (D.S.D. Mar. 5, 1985) (.Buckanaga) [Available on WESTLAW, DCTU database]. For reversal, appellants argue that the district court erred in (1) failing to make detailed findings of fact and in failing to discuss all substantial evidence contrary to its decision, (2) excluding on its own motion a Bureau of Indian Affairs Tribal enrollment list, and (3) finding that there was no violation of the Voting Rights Act. For the reasons discussed below, we reverse and remand.
The District is a public school district organized and governed by South Dakota law. The District encompasses parts of Roberts and Marshall counties in the extreme northeast corner of South Dakota and is primarily rural. Total population in 1980 was 5,628 persons, of which 1,908 (33.9%) were American Indians. The District encompasses a part of the original Lake Traverse Reservation, which was the home of the Sisseton-Wahpeton tribe; the reservation was disestablished in the late 19th century. At the time this action was commenced, 44.9% of the students enrolled in District schools were Indians.
The District board is composed of nine members who are elected on a staggered basis for a three year term. Each board member is elected at large by the voters of the entire District. State law requires that four seats on the board be filled by rural candidates; the other five seats are open to any qualified candidate. No Indian served on the District board until 1968, when an Indian was appointed to the board to serve an unexpired term. Since 1969 one Indian has been a member of the school board. The District is divided into two polling precincts with polling places located in Peever and Sisseton.
Appellants filed this action on May 5, 1984, alleging that the District’s at-large election system unlawfully dilutes Indian voting strength and has a discriminatory effect in violation of § 2 of the Voting Rights Act. Appellants sought implementation of a plan under which all nine school board members would be elected from single-member districts or wards. On June 12,1984, the district court entered a temporary restraining order enjoining the District from conducting the June 19, 1984, District election. Following a consideration of the parties’ briefs and arguments, the district court vacated the temporary restraining order on February 21, 1985. On March 5,1985, the district court entered judgment in favor of the District and held that appellants had failed to prove a viola*471tion of § 2 of the Voting Rights Act or of the fourteenth and fifteenth amendments. This appeal followed.
Section 2 of the Voting Rights Act
In 1982 Congress amended § 2 of the Voting Rights Act to prohibit not only those voting practices directly prohibited by the fifteenth amendment but also any voting practice or procedure “imposed or applied ... 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 color____” 42 U.S.C. § 1973(a). The legislative history of the 1982 amendment to § 2 indicates that it was aimed particularly at discriminatory at-large election systems which dilute minority voting strength. H.R.Rep. No. 227, 97th Cong., 1st Sess. 18 (1981); S.Rep. No. 417, 97th Cong., 2d Sess. 30, 38-39 (1982), reprinted in 1982 U.S. Code Cong. & Ad. News 177, 206-07. See also United States v. Marengo County Comm’n, 731 F.2d 1546,1553 (11th Cir.1984); Jones v. City of Lubbock, 727 F.2d 364, 369 (5th Cir.1984). Minority voters may establish a violation of § 2 if they prove that “based on the totality of the circumstances” they 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 legislative history of § 2 enumerated “typical objective factors” to guide the courts in analyzing the discriminatory result of an election system or practice:
(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’s
slating process, whether the members of the minority group have been denied access to that process; (5) the extent to which the 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 hinders their ability to participate effectively in the political process; (6) whether the political campaigns have been characterized by overt or subtle racial appeals; (7) the extent to which members of a minority group have been elected to public office in the jurisdiction.
S.Rep. No. 417, at 28-29, H.R.Rep. No. 227, at 30, reprinted in 1982 U.S. Code Cong. & Ad. News at 206-07; see Jones v. City of Lubbock, 727 F.2d at 379. Congress also cited two other factors which might have limited relevance.
Whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group. Whether the policy underlying the state or political subdivisions’ use of such voting qualifications, prerequisite to voting, standard, practice or procedure is tenuous.
S. Rep. No. 417, at 29, reprinted in 1982 U.S. Code Cong. & Ad. News at 207. Congress made clear, however, that the factors identified in the legislative history are not exhaustive and that other factors may be indicative of a violation and that there is no formula for aggregating the factors. Id.; United States v. Marengo County Comm’n, 731 F.2d at 1574.
The Supreme Court in a recent case, Thornburg v. Gingles, — U.S. —, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986) (Thornburg), for the first time construed § 2 of the Voting Rights Act. Thornburg involved a challenge to a legislative redistricting plan of multi-member districts in five North Carolina legislative districts. While recognizing that all of the factors listed in the Senate Report may be relevant to a claim of vote dilution, the Court held that the existence of certain circumstances
*472was a necessary precondition to establishing a claim of vote dilution. First, a minority group must be able to show that
as a threshold matter ... it is sufficiently large and geographically compact, to constitute a majority in a single-member district____ Unless 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 the structure or practice. [I]f, 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., 106 S.Ct. at 2766 n. 17. “Second, the minority group must be able to show that it is politically cohesive. If the minority group is not politically cohesive, it cannot be said that the selection of a multimember electoral structure thwarts distinctive minority group interests.” Id. at 2767. “Third, the minority must be able to demonstrate that the white majority votes sufficiently as á block to enable it — in the absence of special circumstances, such as the minority candidate running unopposed ... — usually to defeat the minority’s preferred candidate.” Id. It is against this background that we consider appellants’ contentions.
I.
Appellants initially argue that the district court failed to make detailed findings of fact and failed to discuss all substantial evidence contrary to its decision. We agree. Fed.R.Civ.P. 52(a) requires the district court in court-tried cases to make findings of facts and conclusions of law sufficiently detailed that the court of appeals can ascertain the factual and legal bases for the district court’s ultimate conclusion. The Fifth Circuit Court of Appeals, which has decided the majority of voting rights cases, has repeatedly emphasized the necessity for very detailed findings of facts in voting dilution cases.
Because the resolution of a voting dilution claim requires close analysis of unusually complex factual patterns, and because the decision in such a case has the potential for serious interference with state functions, we have strictly adhered to the Rule 52(a) requirements in voting dilution cases and have required district courts to explain with particularity the reasoning and the subsidiary factual conclusions underlying their reasoning.
Velasquez v. City of Abilene, 725 F.2d 1017, 1020 (5th Cir.1984). The findings of the district court herein are inadequate to permit this court to properly perform its appellate review function. We therefore remand the case for further findings. On remand the district court must discuss not only the evidence that supports its decision but also all the substantial evidence contrary to its opinion as it relates to each of the factors discussed below. See id.
Size and Geographical Distribution of Indian Population
As indicated above, the Supreme Court held in Thornburg that a threshold question in a vote dilution case is whether the minority group is “sufficiently large and geographically compact to constitute a majority in a single member district.” 106 S.Ct. at 2766 n. 17. If the minority group is unable to demonstrate this, as the Court suggested would be the case in “a substantially integrated district, the multi-member form of the district cannot be responsible for minority voters’ inability to elect its candidates.” Id. (emphasis in original; footnote omitted).
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 representa*473tives of their choice in the absence of the multimember electoral structure.
Id. n. 17.
The district court in determining this threshold question did not have the advantage of the Supreme Court’s decision in Thornburg to guide it. On remand the district court should consider and make specific and detailed findings concerning the size and distribution of Indians in the school district and determine if the group is sufficiently large and compact to constitute a majority in a single-member district. Id. at 2767 n. 17.
Racially Polarized Voting: Majority and Minority Bloc Voting
The legislative history of § 2 indicates that the presence of racially polarized voting will ordinarily be the keystone of a vote dilution case. S. Rep. No. 417 at 28-30, reprinted in 1982 U.S. Code Cong. & Ad. News at 206-07; Thornburg, 106 S.Ct. 2769. The surest indication of race conscious politics is a pattern of racially polarized voting extending over time. Thorn-burg, 106 S.Ct. at 2770; United States v. Marengo County Comm’n, 731 F.2d at 1567; Jones v. City of Lubbock, 727 F.2d at 381.
What constitutes racially polarized voting was a major issue in Thornburg. Although the Court recognized that “the extent of bloc voting necessary to demonstrate that a minority’s ability to elect its preferred representative is impaired varies according to several factual situations, [and] the degree of bloc voting which constitutes the threshold of legal significance will vary from district to district,” the Court nonetheless set out several general principles concerning racial bloc voting:
The purpose of inquiring into the existence of racially polarized voting is twofold: to ascertain whether minority group members constitute a politically cohesive unit and to determine whether whites vote sufficiently as a bloc usually to defeat the minority’s preferred candidates. Thus, the question whether a given district experiences legally significant racially polarized voting requires discrete inquiries into minority and white voting practices. A showing that a significant number of minority group members usually vote for the same candidate 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. And, in general, a white bloc vote that normally will defeat the combined strength of minority support plus white “crossover” votes rises to the level of legally significant white bloc voting. The amount of white bloc voting that can generally “minimize or cancel” ... [minority] voters’ ability to elect representatives of their choice, however, will vary from district to district according to a number of factors, including the nature of the allegedly dilutive electoral mechanism; the presence or absence of other potentially dilutive electoral devices, such as majority vote requirements, designated posts, prohibitions against bullet voting; the percentage of registered voters in the district who are members of the minority group; the size of the district; and, in multimember districts, the number of seats open and the number of candidates in the field.
106 S.Ct. at 2769-70 (citations and footnotes omitted).
In the present case, the district court erred in failing to make findings concerning racially polarized voting. There was substantial evidence presented by appellants that voting in the District was polarized along racial lines. In part appellants attempted to establish this factor through the deposition of Dr. Gordon Henderson. Dr. Henderson analyzed district elections from 1976 to 1982, and concluded that there was a pattern of racially polarized voting. To refute appellants’ contention, appellees submitted the deposition of Dr. Robert Burns, who concluded that there was no evidence of polarized voting. The district court generally rejected Dr. Henderson’s testimony because Dr. Burns had more experience in South Dakota. More particularly, the district court stated that it *474could not accept the opinion of Dr. Henderson “as to which candidate any particular Indian voter cast a ballot for in the District election.” Buckanaga, slip op. at 8.
Deposition testimony can be assessed by the district court on the basis of credibility, see Hoefelman v. Conservation Comm’n, 718 F.2d 281, 284-85 (8th Cir. 1983), but the district court may not “insulate its findings from review by denominating them credibility determinations.” Anderson v. City of Bessemer, 470 U.S. 564, 105 S.Ct. 1504, 1512, 84 L.Ed.2d 518 (1985). In order for this court to properly review the district court’s factual findings, the district court must set out more completely thé basis for its credibility determination and how these factors relate to the evidence. In addition, the district court should make clear how its determination regarding the testimonial evidence bears upon its determination of the ultimate factual question of polarized voting.
History of Discrimination
Without a doubt a history of discrimination against a minority is important evidence of both discriminatory intent and discriminatory results. “A history of pervasive, purposeful discrimination may provide strong circumstantial evidence,” United States v. Marengo County Comm’n, 731 F.2d at 1567, (1) that the present day acts of elected officials are motivated by the same purpose, or by a desire to perpetuate the effects of that discrimination, (2) that the present day ability of minorities to participate on an even footing in the political process has been seriously impaired by the past discrimination, and (3) that past discrimination has also led to present socio-economic disadvantages, which in turn reduces participation and influence in political affairs. Id.
The district court found that “there [was] no evidence that South Dakota has a historical background of racial discrimination comparable to the South.” Buckanaga, slip op. at 4. The district court also found that “Indians were permitted to vote in Roberts County during the years referred to in the affidavit of attorney Milton Cameron.” Id. at 8. The district court failed, however, to discuss substantial contrary evidence that South Dakota officially excluded Indians from voting and holding office until the 1940s. Nor did the district court consider that even after the repeal of state laws expressly denying the Indians the right to vote, the State of South Dakota, as late as 1975, effectively denied Indians the right to vote in certain county elections. Little Thunder v. South Dakota, 518 F.2d 1253, 1255-57 (8th Cir.1975). More recently this court held that a denial of an Indian resident’s right to run for the office of county commissioner violated the equal protection clause. United States v. South Dakota, 636 F.2d 241, 244-45 (8th Cir.1980).
Moreover, the district court did not discuss substantial evidence regarding the present social and economic disparities between Indians and whites. Evidence was presented that Indian residents in the District continue to have disproportionately lower educational attainment than whites (46% of Indians in Roberts County graduated from high school compared with 56.4% of whites); disproportionately lower median income ($16,321 for Whites as compared with $9,149 for Indians); significantly higher poverty rates (45.8% of all Indians in South Dakota are below the poverty level, as compared with only 11.5% of whites); an unemployment rate which is more than four times that of whites (21.7% unemployment for Indians contrasted with 5.1% for Whites); and Indians are disproportionately located in the poorest paying, lowest status jobs available. The district court also failed to discuss the disparity between Indian voting registration in South Dakota as compared to white voter registration (30.71% for whites compared with 9.94% for Indians). On remand the district court must make specific findings concerning the contrary evidence of discrimination against Indians and any lingering effects of discrimination as evidenced by economic and social disparity between Indians and whites *475and the effect of this discrimination and disparity upon Indian political participation. Election Practices
The district court found that many of the voting practices and procedures that may be used to enhance the opportunity for discrimination against the minority group were nonexistent in South Dakota. Specifically the court found that there was no precondition to candidacy except the timely filing of a petition signed by 20 eligible voters of the district. Buckanaga, slip op at 5. There was no primary, no slating, no majority requirement and no ban on “single shot” voting. In addition the district court found that the District had never been charged with discouraging any Indian residents from registering to vote or from voting. Id. The district court further found that the distances to the polling places were not unreasonable given the rural area involved and that the District was justified in not establishing more polling places within the reservation without the express concurrence of the tribal council. Id. The district court also found that the increase in the size of the board from five to nine members had the effect of making it more likely that some Indian candidates would be elected. Id.
The district court, however, failed to discuss two election practices that appellants contend have a discriminatory effect. These are the staggered terms and the South Dakota statute that apportions school board seats between rural and urban members on the basis of registered voters. We agree with appellants that the district court should have considered evidence of the discriminatory impact of these practices.
The Supreme Court has held that staggered terms promote the dilution of minority voting strength because they limit the number of seats, create more head-to-head contests between white and minority candidates, which highlight the racial element and minimize the influence of single-shot voting. City of Rome v. United States, 446 U.S. 156, 185 & n. 21, 100 S.Ct. 1548, 1565, 1566 n. 21, 64 L.Ed.2d 119 (1980). Thus, a staggered term requirement combined with a white majority and white bloc voting places a minority at a severe disadvantage. See Perkins v. City of West Helena, 675 F.2d 201, 212 (8th Cir.), aff’d, 459 U.S. 801, 103 S.Ct. 33, 74 L.Ed.2d 47 (1982). The discriminatory effect of staggered terms was specifically considered by Congress in the enactment of § 2. H.Rep. No. 227 at 18; S.Rep. No. 417 at 143-44; see also Jones v. City of Lubbock, 727 F.2d at 383. On remand, the district court must make specific findings concerning the effect of staggered terms on District elections.
The district court also failed to consider and make specific findings concerning appellants’ evidence that the apportionment of seats on the school board based on voter registration diluted Indian voting strength. Appellants argue that the historical discrimination against Indians in South Dakota and the lingering socio-economic effects have resulted in a substantially lower voter registration rate among Indians. Because of this lower voter registration rate and the concentration of Indians in rural areas, the number of seats that are allocated to rural areas has been depressed.
Congress intended that the “vestigial effects of past purposeful discrimination,” Thornburg, 106 S.Ct. at 2776 (plurality opinion), would be eradicated by the Voting Rights Act. Low political participation is one of the effects of past discrimination. “Both [the Supreme Court] and other federal courts have recognized that political participation tends to be depressed where minority group members suffer effects of prior discrimination such as inferior education, poor employment opportunities, and low incomes.” Id. Using voter registration as a basis for apportioning electoral seats in some instances may result in perpetuating the effects of past discrimination against minority groups. The district court must consider if the South Dakota statutory requirement has this effect when combined with other factors.
The district court also did not make specific findings concerning the size of the *476District, the presence of only two polling places in the District and the effect of these practices on Indian voter participation. The district court did state that the alternate plan suggested by appellants would result in “very thinly populated districts” and “could well do all district residents more harm than good.” Buckanaga, slip op. at 5. The district court also found that the distances to the polling places were not unreasonable, and that the District was justified in not establishing a polling place on the reservation absent the express concurrence of the tribe through its council. The district court, however, does not indicate what evidence it considered in arriving at these conclusions. Effective review by this court requires more discussion of the evidence and detailed findings.
Minority Electoral Successes The district court’s findings on Indian electoral success focused only on the successes. The district court noted that for 12 of the past 16 years, an Indian had served on the school board and that during 1973-74, two served. Buckanaga, slip op. at 6. The district court judicially noticed that an American Indian had been elected to the U.S. House of Representatives and served 10 years in the House, and that an American Indian unsuccessfully ran for attorney general of South Dakota in 1970 and came within 2,000 votes of winning the office. Id. at 8-9. The district' court, in considering the absence of Indian electoral candidates and office holders in the cities and county, stated:
The court finds that it is not unusual that no Indians have been candidates for city office in Sisseton or for county office in Roberts County. The court finds that the enrolled tribal Indian quite naturally looks to the tribal government in the way that white residents look to the city and county government.
Id. at 7-8. The district court does not identify what evidence in the record supports this conclusion.
Section 2 provides that “[t]he extent to which members of a protected class have been elected to office ... is one circumstance which may be considered.” 42 U.S.C. § 1973(b). The legislative history, however, expressly states that “the election of a few minority candidates does not ‘necessarily foreclose the possibility of dilution of the black vote ...’” S.Rep. 417 at 29, reprinted in 1982 U.S. Code Cong. & Ad. News at 207.
Such success might, on occasion, be attributable to the work of politicians, who, apprehending that the support of a black candidate would be politically expedient, campaign to insure his [or her] election. Or such success might be attributable to political support motivated by different considerations — namely that election of a [minority] candidate will thwart successful challenges to electoral schemes on dilution grounds. In either situation, a candidate could be elected despite the relative political backwardness of [minority] residents in the electoral district.
Thornburg, 106 S.Ct. at 2780 n. 37, quoting Zimmer v. McKeithen, 485 F.2d 1297, 1307 (5th Cir.1973). Thus, the success of some minority candidates does not foreclose a § 2 claim. Thornburg, 106 S.Ct. at 2780.
The district court must conduct an “independent consideration of the record” and a “searching practical evaluation”- of the circumstances surrounding minority electoral successes. Id. The district court has not conducted this type of analysis nor made the proper detailed findings following such an analysis. The district court failed to consider substantial contrary evidence that the Indian community had not been successful in electing persons to public office. Appellants offered evidence that the nine member school board had been all-white until 1968 despite the fact that the Sisseton School District was 34% Indian; that from 1974 to the present, there has been only one Indian board member; and since 1982, 23 Indians sought office and only 3 were successful. Appellants also offered evidence to suggest that the one Indian school board member was “handpicked” by the all white school board, and *477was elected by the white voters. Brief for Appellants at 37. The district court erred in failing to consider this contrary evidence, and in failing to make specific detailed findings in light of this evidence.
Unresponsiveness
Although appellants presented evidence on the District’s lack of responsiveness to Indian needs, the district court made no findings on this issue. The legislative history of § 2 makes clear that unresponsiveness may be raised at the plaintiff’s discretion, although it is not a necessary factor. S.Rep. No. 417 at 29 n. 16, reprinted in 1982 U.S.Code Cong. & Ad.News at 207; United States v. Marengo County Comm’n, 731 F.2d 1546. Appellants offered substantial evidence that the District was unresponsive to the Indian community: (1) although 44.9 percent of the Sisseton’s School District enrollment is Indian and only 12 of the 195 positions in the school district are filled by Indians and federal law already mandates that 6 of these 12 positions must be filled by Indians, (2) the District has never appointed an Indian to serve as an election or poll official, and (3) the District failed to establish additional polling places despite the request of appellants. The district court may not ignore this evidence. The district court must consider it along with the other evidence of vote dilution offered by appellants and must make findings indicating whether and in what manner this evidence supports or does not support appellants’ voter dilution claim.
II.
Appellants next argue that the district court erred in excluding a Bureau of Indian Affairs tribal enrollment list on the basis that the list was not certified by the Bureau of Indian Affairs.
The identification of minority voters is crucial in a vote dilution case because the minority group must show it is sufficiently large and geographically compact to constitute a majority in a single-member district. Thornburg, 106 S.Ct. at 2766 n. 17. Plaintiffs must identify the minority group and then establish that members of the group are registered in lower numbers, vote in lower numbers or are unable to elect the candidate of their choice in violation of § 2. Appellants attempted to identify Indian voters through two methods. Initially, appellants offered the testimony of two persons who examined the voter registration lists and identified voters as Indians by their surname. The district court questioned this method because of the possibility that the surname was not indicative of Indian descent. In response to this concern, appellants submitted a Bureau of Indian Affairs enrollment list1 of members of the Sisseton-Wahpeton Sioux Tribe. The district court excluded the enrollment list because it was “lacking Bureau of Indian Affairs certification or official tribal certification as to enrollment in the tribe.” Following the exclusion of this list, the district court determined that there was insufficient evidence to establish “accurately the identity of each voter as to Indian or non-Indian.” Buckanaga, slip op. at 8.
Because we are remanding this case on other grounds, appellants will have an opportunity to secure a certified enrollment list and to submit this list to the district court and to appellees. We therefore need not decide whether the district court abused its discretion in excluding this evidence.
III.
Appellants lastly argue that the district court erred in holding that there was no § 2 violation. As we have indicated above, appellate review of the district court’s ultimate holding is not possible at this time because of the absence of sufficiently de*478tailed findings based on all the evidence in the record.
Accordingly, we reverse the judgment of the district court and remand for further proceedings consistent with this opinion. This court will retain appellate jurisdiction.
. The list included tribal members’ names, address, date of birth, and identification number. The enrollment list was cross-referenced to the Sisseton-Wahpeton Sioux Housing Project list, and to the school district registration list to determine the number of Sisseton-Wahpeton Sioux members who are registered to vote.