Plaintiffs appeal the judgment of the district court finding that the Raymondville .Independent School District does not discriminate against Mexican-Americans in its ability grouping and teacher hiring practices, and that the Raymondville Independent School District has implemented an adequate bilingual education program under federal law. For the reasons set forth below, we affirm.
I.
Plaintiffs, Mexican-American children and their parents representing a class of others similarly situated, instituted this action against the Raymondville Independent School District (“RISD”) alleging that the district engaged in policies and practices of racial discrimination against Mexican-Americans depriving Plaintiffs and their class of rights in violation of the fourteenth amendment and 42 U.S.C. § 1983 (1976), Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d et seq. (1976), and the Equal Educational Opportunities Act of 1974, 20 U.S.C. § 1701 et seq.' (1976) (“EEOA”). Plaintiffs claimed that the school district unlawfully discriminated against them by using an ability grouping system for classroom assignment which was based on racially and ethnically discriminatory criteria and resulted in impermissible classroom segregation; by discriminating against Mexican-Americans in the . hiring and promotion of teachers and administrators; and by failing .to imple*459ment adequate bilingual education to overcome linguistic barriers impeding Plaintiffs' equal participation in the educational program of the district. The factual and procedural history of this litigation is set forth in our earlier opinion, Castaneda v. Pickard, 648 F.2d 989 (5th Cir.1981), and we see no need to repeat that history here.1
In Castaneda I, we affirmed the district court’s conclusion that RISD’s bilingual education program is not violative of Title VI. However, we reversed and remanded the district court’s findings for RISD on the other issues appealed. We instructed the district court to inquire into the history of RISD to determine whether RISD had in the past discriminated against Mexiean-Americans, and to consider whether the vestiges of such past discrimination had been erased. The answers to these questions would affect the analysis of Plaintiffs’ claims that the ability grouping and employment practices of RISD are tainted by unlawful discrimination. We further directed the district court to conduct a hearing to address the question of the legality of the district’s language remediation program under the EEOA, 20 U.S.C. § 1708(f).
On remand, Plaintiffs amended their complaint to name the Texas Education Agency (“TEA”) as a defendant, alleging that the TEA had failed to fulfill its duty to assure that Plaintiffs were not subjected to discriminatory practices through the use of state or federal funds and that an adequate bilingual program was implemented by RISD. After a second trial, the district court determined that no vestiges of discrimination remained in RISD and that the ability grouping system of RISD was not discriminatory. The district court also found RISD’s recruiting and employment practices to be free from discrimination. Finally, the district court held that RISD’s bilingual education program survived scrutiny under the EEOA.
On appeal, Plaintiffs claim error in the district court’s conclusions that there are no remaining vestiges of historical discrimination in RISD; that RISD rebutted Plaintiffs’ prima facie case of discrimination in the hiring and promotion of Mexican-Americans; and that RISD and the TEA have taken appropriate action and made genuine efforts to overcome the language disabilities of RISD students. We do not agree that the district court erred in its conclusions. While we remain committed to Plaintiffs’ right to be free from discrimination and their right to receive an education comparable to that given Anglo-American students, the record before us today, revealing a wide array of changes made in RISD and Texas state law since the date of the first trial below, does not mandate a conclusion that these rights have been violated. After a thorough review of the record compiled from the second trial in this case,2 we affirm.
II.
In Castaneda I, we determined that Plaintiffs’ challenge to the ability grouping practices of RISD was hampered by the district court’s failure to make findings on whether there had been past discrimination by any RISD and whether there remained any lingering vestiges of discrimination within the school district. We observed that
[i]f the district court finds that RISD has a past history of discrimination and has not yet maintained a unitary school system for a sufficient period of time that the effect of this history may reasonably be deemed to have been fully erased, the district’s current practices of ability grouping are barred....
Castaneda I, 648 F.2d at 997. Conversely, if RISD were found to have no history of discrimination, or
if despite such a history, the system has achieved unitary status and maintained such status for a sufficient period of *460time that it seems reasonable to assume that any racially disparate impact of the ability grouping does not reflect either the lingering effects of past segregation or a contemporary segregative intent, then no impermissible racial classification is involved and ability grouping may be employed despite segregative effects.
Id. at 996. We further instructed the district court to examine whether the ability grouping practices operated to confuse measures of language and intelligence; if so, the ability grouping could itself be evidence of a discriminatory intent. Id. at 998.
The district court accordingly looked to indicia of past discrimination and remaining vestiges of same and concluded that “RISD has been a unitary system for a sufficient period of time that any vestiges of discrimination have been erased.” Record Vol. 1 at 12. The court noted that with the implementation of a freedom-of-choice attendance policy in 1972, students were provided with the opportunity to attend the school of their choice. The ethnic composition of students at L.C. Smith Elementary School had shifted from 100% Mexican-American in 1971-72, to 97.88% Mexican-American, 2.12% Anglo-American in 1983-84.3 Castaneda II, Plaintiffs’ Exhibit 57. The school board was composed of three Mexican-Amerieans and four Anglo-Americans. The court observed that since “vestiges of discrimination have been erased,” the evidence of ability grouping should be reviewed “without the indicia of historical discrimination.” Record Vol. 1 at 12. The court held that RISD’s ability grouping practices did not mandate a conclusion of discrimination.
Plaintiffs argue that the district court erred in finding that there are no remaining vestiges of historical discrimination in RISD. That vestiges remain can be deduced from the high percentage of Mexican-American students attending L.C. Smith and the failure of the freedom-of-choice plan, Plaintiffs contend. Most importantly, Plaintiffs claim that the ability grouping system of RISD is itself a vestige of past discrimination: the district court’s refusal to so find is clearly erroneous.4
Our review here of the district court’s conclusion that “RISD has been a unitary system for a sufficient period of time that all vestiges have been erased” is controlled by Federal Rule of Civil Procedure 52, which provides that findings of fact “shall not be set aside unless clearly erroneous.” Fed.R.Civ.P. 52(a). United States v. Texas Education Agency, 647 F.2d 504, 506 (5th Cir.1981), cert. denied, 454 U.S. 1143, 102 S.Ct. 1002, 71 L.Ed.2d 295 (1982) (holding clearly erroneous the district court’s finding that a 1970 plan had eliminated all vestiges of the previous dual system and created a unitary school); Calhoun v. Cook, 522 F.2d 717 (5th Cir.1975) (holding not clearly erroneous the district court’s findings that the Atlanta School District was unitary and that the remaining one race schools were not vestiges of past discrimination). See also Tasby v. Wright, 713 F.2d 90, 93 (5th Cir.1983) (observing that despite upholding the district court’s determinations of unitary status as not clearly erroneous, in two previous cases, the court would not find the district court’s failure to declare unitary status in the instant case clearly erroneous). The clearly erroneous standard bids us to affirm the district court’s findings unless, upon review of the record as a whole, we become convinced that a mistake has been committed. United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948). As the Supreme Court has recently noted, “if the district *461court’s account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact it would have weighed the evidence differently.” Anderson v. City of Bessemer, - U.S. -, 105 S.Ct. 1504, 1512, 84 L.Ed.2d 518 (1985). If there are “two permissible views of the evidence, the fact finder’s choice between them cannot be clearly erroneous.” Id.
Public school officials have a continuing duty to eliminate the system-wide effects of earlier discrimination and to create a unitary school system untainted by the past. Swann v. Charlotte-Mecklenburg Board of Education, 402 U.S. 1, 15, 91 S.Ct. 1267, 1275, 28 L.Ed.2d 554 (1970). In determining whether the past has been effectively eradicated insofar as it remains in the power of school officials and courts to do, “we must keep in mind that each school district is unique.” Tasby v. Wright, 713 F.2d at 93 (quoting Ross v. Eckels, 434 F.2d 1140, 1147 (5th Cir.1970), cert. denied, 402 U.S. 953, 91 S.Ct. 1614, 29 L.Ed.2d 123 (1971)). The “determination of remedies turns on conditions in a particular district,” Ross v. Houston Independent School District, 699 F.2d 218, 227 (5th Cir.1983), and in “like fashion, the decision that public officials have satisfied their responsibility to eradicate segregation and its vestiges must be based on conditions in the district, the accomplishments to date, and the feasibility of further measures.” Id. We now turn to consider the particular situation in RISD examined by the district court.
The high percentage of Mexican-American students attending L.C. Smith at the time of the second trial below, 97.88%, does not necessarily mandate a finding that unremedied vestiges of discrimination yet exist. The Constitution does not after all require “that every school in every community must always reflect the racial composition of the school district as a whole.” Swann, 402 U.S. at 24, 91 S.Ct. at 1280. More important, this case presents a situation different from most school cases in that Mexican-Amerieans make up considerably more than a majority of the students in RISD. A poor district in the Rio Grande Valley, Raymondville’s population is estimated to be 77% Mexican-American, with almost all of the remaining 23% Anglo-American. Mexican-American students constitute at least 88% of the student population of RISD. Castaneda II, Plaintiffs’ Exhibit 57. Anglo-American students, the majority in most school cases, here represent a mere 12% of the student population. Any chance of true desegregation in this school district would therefore appear impossible. The situation in RISD seems similar to that we addressed in Calhoun v. Cook, 522 F.2d 717 (5th Cir.1975). In Calhoun, we observed that given the fact that 85% of the students in the Atlanta, Georgia, school district were black, the prospects for any intradistrict desegregation of the city schools were bleak. We accordingly upheld the district court’s finding that Atlanta’s remaining one-race schools were the product of its preponderant majority of black pupils rather than a vestige of past discrimination. Calhoun, 522 F.2d at 719. Similarly, in Ross v. Houston Independent School District, 699 F.2d 218 (5th Cir.1983), we upheld as not clearly erroneous the district court’s finding that the Houston school system was unitary despite a student population at several schools that was heavily Hispanic and black. Here too we cannot find the district court clearly erroneous in failing to find the high percentage of Mexican-American students attending L.C. Smith a vestige of discrimination.
The district court’s reliance on the freedom-of-choice plan in finding RISD unitary also does not seem in error. The record reveals that eleven Anglo-American students currently choose to attend L.C. Smith. Castaneda II, Plaintiffs’ Exhibit 57. We have in the past found it unnecessary when a freedom-of-choice plan is in effect to divide a remaining small number of whites, already in a minority position, amongst schools. See Singleton v. Jackson Municipal Separate School District, *462419 F.2d 1211, 1221 (5th Cir.1969), rev’d in part sub. nom., Carter v. West Feliciana School Board, 396 U.S. 290, 90 S.Ct. 608, 24 L.Ed.2d 530 (1970).5 Of course, Plaintiffs here have not requested that we do so. In fact, Plaintiffs here, as in Ross, have not come up with any proposal that would further desegregate RISD. In Plaintiffs’ failure to suggest further means of desegregation we see a tacit recognition that what has been done, even if not perfect, cannot be termed a vestige of discrimination. In sum, we affirm the district court’s finding that the benign policies of classroom assignment as augmented by the freedom-of-choice plan are maintaining RISD as a unitary school system.
Plaintiffs next seek review of the district court’s failure to find vestiges of discrimination in the ability grouping of students in RISD. In ruling that RISD’s ability grouping practices were not discriminatory, the district court implicitly found that the ability grouping practices were not themselves vestiges' of discrimination and that the ability grouping did not improperly confuse language and intelligence. Plaintiffs now argue that both these implicit findings are clearly erroneous, that RISD’s ability grouping practices continue to discriminate against Mexican-Americans by relying upon English achievement tests as “the decisive factor” in placing students in ability groups.
Plaintiffs’ description of RISD’s use of English achievement tests contrasts dramatically with that adopted by the district court. We find that the district court’s observations on the use of English achievement tests in RISD are supported by the record. At trial, Joe Herod, Curriculum Director for RISD, elaborated upon the district’s ability grouping practices. Record Vol. 12 at 478-85. In early elementary grades, a program called “The Early Prevention of School Failures” is in effect. A battery of five tests is given to each student entering kindergarten. These tests are given in the language that the child best understands, English or Spanish. The tests include a language assessment test, a visual motor integration test, a “draw a person” test, and a “Peabody Picture” test. All tests are “developed to assess the maturity of each child as pertaining to their readiness to start academic-type programs in school.” Record Vol. 12 at 479. Based on the results of this battery of tests, the students are assigned to a “no risk,” “moderate risk” or “high risk” group defining their readiness to learn reading skills. Record Vol. 12 at 480.
In the first grade, criterion referenced tests are administered in English or Spanish. Unlike achievement tests, which attempt to discover how the test-taker compares to the typical or average person, criterion referenced tests have a predetermined standard that must be met. Here, the criterion referenced test indicates which of the students have achieved first grade educational objectives. This test is one factor in the placement of students in the second grade. Record Vol. 12 at 481-82.
In the second grade, for the first time, an English language standardized achievement test is administered during the first week? of class in order to establish pre-test scores for evaluation taking place at the end of the year. Record Vol. 12 at 482. The test is administered only to students who read and write English. Record Vol. 12 at 483. No grouping in the second grade is based on the English language standardized achievement test. Id.
As explained by Mr. Herod and other witnesses, ability grouping in grades three through eight is based on English achievement test scores, school grades, teacher evaluations and the recommendations of school counselors. Test scores are but one factor used in assigning students to one of the three ability groups. Because there is a preference for assigning students to a higher group, achievement scores may be weighed in the calculus as a secondary factor. For example, if achievement scores *463are low, but grades high, less weight will be accorded to the achievement scores. Record Yol. 12 at 484. In addition, a Spanish language achievement test is in the process of design and development.
Plaintiffs claim that the depositions of Mr. Herod, Mr. Jerry Jacobs, Superintendent of RISD, and Mrs. Dhelma Anzaldua Jetton, the Director of Curriculum and the Bilingual Program under Mr. Herod, support their contention that English achievement tests are the “decisive factor” in ability grouping. The cited depositions do not, however, add strength to Plaintiffs’ contention. In fact, Mr. Herod in his testimony directly contradicts Plaintiffs’ characterization of the use of English achievement tests by explaining that the English achievement test is but one factor considered in ability grouping. Deposition of Herod at 38. Mr. Herod observes that teacher recommendations and teacher given grades are also important. Id. at 39. Plaintiffs’ reliance on Mr. Jacobs’ deposition is particularly mystifying, as Mr. Jacobs in his deposition does not discuss the use of English achievement tests at all except insofar as they had served as a model in developing Spanish language achievement tests. Deposition of Jacobs at 53-55. Mrs. Jetton testified that various factors are considered in exiting students from the bilingual program and placing them in the regular program, and that English achievement tests and recommendations are used in grouping students in the regular program. Deposition of Anzaldua, 1982, at 22, 57; Deposition of Jetton, 1983, at 23-24.
Plaintiffs also claim that the expert testimony of Dr. Ernest M. Bernal, a research associate at the University of Texas working in bilingual special education, “confirmed this court’s suspicion that the RISD over-reliance on English language achievement tests for ability grouping may be highly suspect.” Appellants’ Brief at 11. Here too, Plaintiffs’ claim is not supported by the record. Dr. Bernal did testify that he thought “achievement tests alone [are] probably not sufficient for ability grouping.” Record Vol. 10 at 157. He never said RISD overly relied on the tests, however, nor did he in fact discuss RISD’s specific use of the tests at all. After reviewing the record as a whole, with little help given us by Plaintiffs’ misleading citations, we do not find overselves convinced that a mistake has been committed by the district court. The district court was thus not clearly erroneous in failing to find RISD’s ability grouping practices a vestige of discrimination.6
III.
We next address Plaintiffs’ charge that the district court erred in finding RISD’s employment practices nondiscriminatory. In Castaneda I, we observed that the claim of unlawful discrimination against Mexican-Americans in the hiring or promotion of faculty and administrators should be analyzed under the approach used to assess the merits of traditional class action and pattern and practice employment discrimination suits. Castaneda I, 648 F.2d at 1001. Class action or pattern and practice employment discrimination suits require a statistical comparison between the racial composition of the employer’s work force and that of the relevant labor market before a finding of discrimination can be made. In Hazelwood School District v. United States, 433 U.S. 299, 97 *464S.Ct. 2736, 53 L.Ed.2d 768 (1977), the Supreme Court held that the proper comparison in a case involving school teachers was
between the racial composition of [the district’s] teaching staff and the racial composition of the qualified public school teacher population in the relevant labor market.
Hazelwood, 433 U.S. at 308, 97 S.Ct. at 2742. Because the district court here had made no such comparison, we remanded the issue of discrimination in teacher hiring for further findings in accordance with Ha-zelwood.
On remand, the district court determined that the relevant labor market was comprised of teachers graduating from Pan American University at Edinburg, Pan American University at Brownsville, Texas A & I University at Kingsville, Texas A & I University at Corpus Christi, and Laredo State University. These universities are located in counties correlating geographically to the TEA’S Regions I and II.7 Record Vol. 1 at 17. The court acknowledged that from time to time south Texas schools must seek teachers outside Regions I and II, but concluded that the record did not support finding a relevant labor market which stretched beyond these two regions.
Although Plaintiffs on appeal urge that the relevant labor market is located within a 120 degree 40 mile west and south radius of Raymondville, an area which includes Region I, Plaintiffs do not seem to contest directly the district court’s determination of the relevant labor market. In fact, Plaintiffs observe that the district court’s reliance on Regions I and II “is the relevant labor market which the Plaintiffs proved both through statistical evidence and through the testimony of school superintendents and the former Director of Personnel of Pan American University and the Director of Personnel for Texas A & I University.” Appellants’ Brief at 14 n. 6 (emphasis in original). Plaintiffs also admit that they “introduced statistics which included Region II further to the north which might provide the defendant RISD with the outer limits of reasonable recruitment based on geographical proximity.” Id. at 15.
In any event, even if we were to interpret Plaintiffs to dispute the district court’s determination of the relevant labor market, they would not prevail on this claim. The determination of a relevant labor market is one of fact, reviewable only for clear error. Markey v. Tenneco Oil Co., 707 F.2d 172, 174 (5th Cir.1983) (citing Pullman-Standard v. Swint, 456 U.S. 273, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982)). There is ample support in the record for finding that Pan American at Edinburg, Pan American at Brownsville, Texas A & I at Kingsville, Texas A & I at Corpus Christi and Laredo State University constitute the primary source of teachers for south Texas schools. Richard Light, Assistant Superintendent for Personnel at Pharr San Juan Alamo Independent School District, testified that he recruits most of his teachers from Pan American University, Texas A & I and Kingsville. Record Vol. 10 at 8. Marcario Garza, Assistant Superintendent of Personnel at LaJoya Independent School District, also recruits at Pan American, Texas A & I at Corpus Christi, and Texas A & I at Kingsville. Record Vol. 10 at 49-50. Mr. Garza testified that he sends recruiters to Pan American at Brownsville, although he so far has not experienced great success in hiring there. Record Vol. 10 at 49, 57. Mr. Garza also recruits at Laredo State University. Record Vol. 10 at 50, 61. Although Joe Sanchez, Superintendent of Progresso School District, testified that he fulfills almost all his recruiting needs at Pan American at Edinburg, Record Vol. 10 at 36, located within Region I, his testimony does not compel a finding that Pan American and Region I alone should be seen as the relevant labor market. If two permissible views of the evidence are possible, a choice between them cannot be clearly erroneous. United States v. Yellow Cab Co., 338 U.S. *465338, 70 S.Ct. 177, 94 L.Ed. 150 (1949). The district court’s determination of the relevant labor market should therefore stand.8
After defining the relevant labor market, the district court next considered Plaintiffs’ claims under 42 U.S.C. § 1983 (1976),9 the Equal Educational Opportunity Act of 1974, 20 U.S.C. § 1701 et seq. (1976) (“EEOA”),10 and Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d et seq.11 The court invoked a class action or pattern and practice model of proof, based on the model set forth by the Supreme Court in International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). As *466the district court noted, under the Teamsters model, Plaintiffs’ burden was to establish a prima facie case that a policy of unlawful discrimination existed. The burden then shifted to the employer, RISD, to defeat the prima facie showing of a pattern or practice by demonstrating that Plaintiffs’ proof was either inaccurate or insignificant. See Teamsters, 431 U.S. at 360, 97 S.Ct. at 1867. The employer’s defense “must, of course, be designed to meet the prima facie case.” The employer’s burden may “provide a nondiscriminatory explanation for the apparently discriminatory result.” 12 Id. at 361 n. 46, 97 S.Ct. at 1867 n. 46. Any “nondiscriminatory justification offered by the [employer] will be subject to further evidence by [Plaintiffs] that the purported reason for an applicant’s rejection was in fact a pretext for unlawful discrimination.” Id. at 362 n. 50, 97 S.Ct. at 1868.
Plaintiffs proved an available labor pool of 55.33% Mexican-American teachers in the relevant labor market. The Mexican-American teachers employed by RISD made up only 39.29% of the faculty. Castaneda II, Plaintiffs’ Exhibit 39. The district court deemed' the 16.04% statistical differential sufficient to establish a prima facie case of unlawful discrimination. See Teamsters, 431 U.S. at 339, 97 S.Ct. at 1856 (discussing the importance of statistics in proving a prima facie case of employment discrimination). The district court also found, however, that RISD had successfully rebutted Plaintiffs’ prima facie case. On appeal, Plaintiffs argue that the district court erred in crediting RISD’s rebuttal testimony to find no discrimination. Plaintiffs’ burden here is a heavy one. We will affirm the district court’s acceptance or rejection of rebuttal evidence unless clearly erroneous. Payne v. Travenol Laboratories Inc., 673 F.2d 798, 823 (5th Cir.), cert. denied, 459 U.S. 1038, 103 S.Ct. 451, 74 L.Ed.2d 605 (1982) (finding that the district court’s rejection of rebuttal evidence was not clearly erroneous). The district court’s finding of nondiscrimination is also factual and must be reviewed under the clearly erroneous standard. Pullman-Standard v. Swint, 456 U.S. 273, 289, 102 S.Ct. 1781, 1790, 72 L.Ed.2d 66 (1982). As noted in Part II, under the clearly erroneous standard we must accept the district court’s factual findings if they are “plausible.” If such findings are “plausible,” we cannot reverse even though we are “convinced that had [we] been sitting as the trier of fact, [we] would have weighed the evidence differently.” Anderson, 105 S.Ct. at 1512.
Keeping the above principles in mind, we turn to a review of the district court’s reasoning and the underlying record. The district court relied mainly on two arguments advanced by RISD: (1) that RISD is hindered in its recruiting because several factors render RISD a less desirable place in which to work than are its neighboring school districts, and (2) that RISD engages in substantial recruiting efforts. Before proceeding further, we must initially address an objection Plaintiffs raise to the first argument advanced by RISD. Plaintiffs contend that RISD’s reliance on its undesirable living and teaching conditions is “counter-intuitive” here: why would RISD’s disadvantages discourage Mexican-American teachers while inexplicably attracting Anglo-American teachers? RISD need not, however, be characterized as maintaining that its poor living and teaching conditions provide a lure for Anglo-American applicants. As the district court observed, the statewide demand for bilingual teachers, Record Vol. 11 at 331-32, places a high premium on native-speaking Mexican-American applicants. Yet the *467applicants is low. Only 11.425% of the teachers in Texas are Mexican-American. Of this 11.425%, 47% are found in the geographical area corresponding to Regions I and II. Castaneda II, Defendant’s Exhibit 5. The high demand is thus coupled with a low supply, and much of the supply is located in the labor market in which RISD must compete. It is therefore not “counter-intuitive” to argue that RISD’s efforts to recruit the few highly sought after Mexican-American applicants in the State of Texas may be hampered by its less desirable teaching conditions.
We now must turn to an examination of the undesirable factors upon which RISD relied. As the district court noted, testimony at trial revealed a number of factors which influence the employment decisions of teachers in Regions I and II. Salary, fringe benefits, family ties, working conditions, entertainment and housing are all critical factors. Plaintiffs’ expert witness testified that Plaintiffs’ statistical prima facie case was prepared under the assumption that the factors affecting desirability of employment were equally present in all given school districts. Record Vol. 10 at 147. To rebut the prima facie case, RISD endeavored to show that the factors affecting desirability of employment were not present to the same degree in all districts, but that, to the contrary, RISD in many ways is a less desirable place to work than are its neighbors. A rural community far from major metropolitan centers, Ray-mondville offers poor housing and entertainment. Jerry Jacobs, Superintendent of RISD, testified that RISD pays only $900.00 above the state minimum salary, less than the other districts discussed at trial. Record Vol. 12 at 528. RISD hires teachers on a year to year basis: no continuing contract system is used. Record Vol. 12 at 527. Although RISD contributes $27.50 a month toward the health insurance of its faculty, no contribution is so made toward life insurance. Record Vol. 12 at 528. No shuttle bus transportation is provided for teachers. Record at 528. The classrooms on the whole lack air conditioning. Record Vol. 12 at 529.
In order to prove RISD’s reliance on its undesirable factors pretextual, Plaintiffs on appeal discuss working and living conditions in several south Texas school districts employing relatively high percentages of Mexican-American teachers. Despite Plaintiffs’ conclusion that bilingual Mexican-American teachers flock to school districts “large and small, poor and relatively well off, isolated and prosperous,” Appellants’ Brief at 23, Plaintiffs fail to convince us that the district court’s acceptance of RISD’s rebuttal evidence was clearly erroneous.
Plaintiffs quote the testimony of representatives from three school districts in the lower Rio Grande Valley: LaJoya School District, Pharr San Juan Alamo School District, and Progresso School District.13 Marcario Garza, Director of Personnel for the LaJoya School District, testified that of the 300 teachers employed in the district, 70-80% are Mexican-American. Entertainment in LaJoya consists of a local football team and a state park. The beach is ninety minutes from town. Record Vol. 10 at 55. Yet the record indicates that the LaJoya School District offers advantages to prospective teachers well beyond the paltry few mentioned by Plaintiffs. LaJoya pays $2300.00 above the state minimum scale. The high pay significantly helps recruiting. Record Vol. 10 at 52, 57. A shuttle bus is provided to transport teachers living in McAllen and Edinburg into LaJoya School District. Record Vol. 10 at 53. The shuttle bus allows teachers to live in the more *468populated town of McAllen where social activities take place. Record Vol. 10 at 56. Mr. Garza testified that his shuttle bus was “one of the things that helps us” in recruiting teachers. Record Vol. 10 at 55. LaJo-ya School District also offers teachers a continuing contract after three years of teaching. Under a continuing contract teachers are not reelected each year, but instead “remain with the school district as long as they did a good job.” Record Vol. 10 at 68. LaJoya School District also provides free life insurance for its staff and pays for one half of the health insurance premiums. Record Vol. 10 at 68.
Richard Light of Pharr San Juan Alamo School District testified that 80-85% of his teachers are Mexican-American, many hired from nearby Pan American University, Record Vol. 10 at 15, 29, which graduates a high number of Mexican-American teachers. Record Vol. 11 at 333-34. Many Pharr San Juan Alamo teachers commute to Pharr San Juan Alamo from Edinburg, McAllen, and Weslaco. Record Vol. 10 at 14, 18, 20. Pharr San Juan Alamo School District, however, also enjoys recruiting advantages denied RISD. In the first place, Pharr San Juan Alamo is located 3 miles from Edinburg where Pan American University is located, only 8 miles from the University itself. Record Vol. 10 at 14. Second, Pharr San Juan Alamo runs a program for student teachers drawn mainly from Pan American University. Record Vol. 10 at 12. Pharr San Juan Alamo hires 40% of its elementary school teachers from its student teacher program, and Mr. Light admitted that the student teacher program gives Pharr San Juan Alamo an advantage over other districts in hiring teachers from Pan American. Record Vol. 10 at 13.
Finally, Joe Sanchez of the Progresso school system testified that he hires 75% Mexican-American teachers. His district is poor and without significant entertainment or housing. Record Vol. 10 at 37. Here again, however, Plaintiffs fail to establish that Progresso school district is directly comparable to RISD. In the first place, Progresso is a smaller school system than is RISD. Progresso employs forty-eight teachers. RISD, on the other hand, has employed from 190 to 196 teachers in each of the years from 1978 to 1983. As noted above, there is a small number of Mexican-American teachers available in Texas. RISD must therefore hire a much greater number of these teachers than must Progresso in order to achieve a similar percentage of Mexican-American teachers employed. Once we consider the numbers involved, the comparison between the two districts is immediately rendered less convincing. Second, Progresso is located closer to major metropolitan areas than is Raymondville. While Raymondville is 51 miles northeast of McAllen, Progresso is only 24 miles from McAllen, Record Vol. 10 at 41, where, indeed, some of its teachers reside. Third, Progresso pays $1000.00 over the state minimum pay scale, Record Vol. 10 at 38, which compares favorably to the $900.00 RISD pays. In sum, Plaintiffs’ discussion of the three above school districts does not render clearly erroneous the district court’s finding that RISD’s inability to hire more Mexican-Americans stems from its undesirable living and working conditions, and not from discrimination.14
Nor does the district court appear clearly erroneous in its conclusion that “recruiting efforts by RISD are substantial in scope and quantity.” Testimony at trial showed that RISD mails vacancy notices in the spring and summer to a large number of schools, and sends on-campus recruiters to various Texas schools, interviewing at Texas A & I and Pan American University at Edinburg and Brownsville. Record Vol. 12 at 503-06. Plaintiffs assert that most recruitment is done by “Anglo-surnamed” personnel. Superintendent Jacobs, how*469ever, testified that Mrs. Camargo and Mrs. Jetton, Director of Curriculum and the Bilingual Program at RISD, often recruit at Texas A & I. Record Vol. 12 at 505. Mrs. Jetton, although perhaps now technically “Anglo-surnamed,” was born Dhelma An-zaldua: Mrs. Camargo and Mrs. Jetton thus would not seem best characterized as “Anglo-surnamed.” In addition, the Director of Student placement at Texas A & I, Johnny L. Johnson, noted that Mexican-Americans had recruited for RISD at Texas A & I. Record Vol. 11 at 350. Romulo Martinez, former placement director of Pan American University, merely could not remember whether or not RISD sent Mexican-Amerieans to recruit at Pan American University. Record Vol. 11 at 333.
RISD mails recruiting letters directly to students graduating in high demand fields such as bilingual education. Record Vol. 12 at 506. The district sends applications and letters to all teachers making inquiries. Although Plaintiffs complain that RISD fails to keep accurate records, the record again does not bear out this contention. The trial transcript indicates that RISD retains all letters mailed to placement offices and to applicants, and is able to compute the number of applications produced by such letters, and the number of applicants ultimately hired from the pool. Record Vol. 12 at 539-40. When asked what else RISD could do to attract Pan American students, Mr. Martinez ventured only, “pay more money. Its hard to pay more money. Or furnish housing.” Record Vol. 11 at 329. Mr. Johnson voiced a similar opinion when asked what RISD could do to recruit more A & I students: “other than get their salary up a thousand or so or two thousand above the schools within that geographic area. I can’t think of anything.” Record Vol. 11 at 349. We therefore do not find the district court clearly erroneous in concluding that RISD engages in substantial recruiting for new teachers without regard to ethnic origin.
Finally, Plaintiffs contend that the district court failed to consider evidence of individual cases of discrimination which would have given probative weight to the Plaintiffs’ statistical proof of discrimination on a class-wide basis. Since the district court upheld Plaintiffs’ prima facie case, we do not find error in its failure to address each individual claim in order to bolster Plaintiffs’ prima facie case. We observe, in addition, that the district court’s first opinion in this ease did examine the individual claims and found them lacking; thus, it is doubtful that these claims would have given “probative weight” to the statistical case in any event. Moreover, the district court’s failure to discuss explicitly all the facts in the record that Plaintiffs feel supported their claim of discrimination is not clearly erroneous. We may reverse a district court’s finding under the clearly erroneous standard only if, after reviewing the record as a whole, we are left with the definite and firm conviction that a mistake has been made. United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948). This requirement that we review the record as a whole precludes any argument that grounds for reversal can be found by merely pointing out facts which the court did not explicitly discuss even if these facts cut against the district court’s interpretation. Fowler v. Blue Bell, Inc., 737 F.2d 1007, 1013 (11th Cir.1984). After reviewing the record as a whole, we are satisfied that the district court was not clearly erroneous in finding that RISD had rebutted Plaintiffs’ prima facie case and in concluding that RISD does not discriminate against Mexican-Americans in hiring teachers.15
*470IV.
The district court on remand found that most of the bilingual teachers in Raymond-ville were fluent in Spanish, and that the teacher training and evaluation procedures were in compliance with state law and the EEOA. On appeal, Plaintiffs claim that the district court erred in concluding that RISD and the TEA had taken “appropriate action” and made genuine efforts to overcome the language disabilities of RISD students in accordance with § 1703(f) of the EEOA. Section 1703(f) makes it unlawful for an educational agency to fail to take “appropriate action to overcome language barriers that impede equal participation by its students in its instructional programs.” Plaintiffs ask that the TEA provide expert assistance to the school district to remedy the deficiencies in the RISD language program, particularly the Spanish language capabilities of the teachers and in-service training.16 Plaintiffs ask us to further require the TEA to monitor RISD on a continuing basis to ensure compliance with our mandate in all respects.
Although Plaintiffs framed their complaint in terms of a § 1703(f) violation, on remand they took the position before the district court that the appropriate program to be implemented by RISD under § 1703(f) is the bilingual program provided by state law. In their brief upon remand below, Plaintiffs informed the court that an important aspect of our decision in Castaneda I had been modified by the subsequent passage by the Texas Legislature in 1981 of the Texas Bilingual Education Act, Tex.Educ.Code § 21.451 et seq. (Vernon Supp.1985), and its rules and regulations. The Act is popularly known as Senate Bill 477. Record Vol. 1 at 117. Plaintiffs “submit[ted]” that the state law “is the appropriate standard against which to judge the RISD program in terms of its appropriateness for purposes of the EEOA, 20 U.S.C. § 1703(f).” Record Vol. 1 at 118. The district court accordingly examined RISD’s bilingual program under the requirements of state law and concluded that as RISD was in compliance with state law, it passed muster under § 1703(f). Because of Plaintiffs’ position before the district court, our review here must center on RISD’s compliance with state law. We by no means imply, however, that a state must provide a program of bilingual education to all limited English speaking students in order to satisfy § 1703(f) of the EEOA. We hold fast to our conviction, voiced in Castaneda I, that in enacting § 1703(f), Congress intended to leave state and local educational authorities a substantial amount of latitude in choosing the programs and techniques they would use. Castaneda I, 648 F.2d at 1009. If Plaintiffs had not taken the position before the district court that they did here, our review of a school district’s compliance with § 1703(f) would be different.
At the time of the first trial below, only one-half of the teachers employed in the bilingual education program at RISD were Mexican-American and native Spanish speakers. Castaneda I, 648 F.2d at 1005. The other teachers in the program received their knowledge of Spanish (a 700 word vocabulary) and their understanding of the theories and methods of bilingual education from a 100 hour course designed by the TEA. Because the record indicated that “some of the teachers employed in the *471RISD bilingual program [had] a very limited command of Spanish,” we entertained “serious doubts about the actual language competency of the teachers employed in bilingual classrooms by RISD.” Id. at 1012-13. Therefore we instructed the district court on remand to attempt to identify more precisely the cause or causes of the Spanish language deficiencies experienced by some of the RISD teachers and to require an improved in-service training and evaluation program. Id. at 1013.
On remand, the district court observed that the facts had changed significantly since the record was compiled in the first trial. The district court found that of the twenty-seven teachers employed by RISD in the 1983-84 school year, sixteen had been hired since the date of the original trial in June 1978. The district court found that twenty-three out of the twenty-seven teachers employed by RISD in 1983-84 were native speakers of the Spanish language, contrasting dramatically with our observations based on the 1978 record in Castaneda I. The remaining four teachers were “proficient in Spanish.” Plaintiffs claim that the district court was clearly erroneous in so finding, and allege that the record indicates many RISD teachers “who readily admitted their own language limitations and inability to use Spanish beyond a kindergarten level.” Plaintiffs’ Brief at 34. RISD must evaluate these teachers, Plaintiffs argue.
We do not agree. Our review of the record from the second trial below leads us to the conclusion that the district court was not clearly erroneous in finding the current RISD teachers to be fluent or proficient in Spanish. Jerry Jacobs, Superintendent of RISD, testified to the Spanish language capabilities of each teacher. Record Vol. 12 at 517-21. Mr. Raymond Magallanes, Education Program Director of the Bilingual Education division at the TEA, visited at least fourteen of the bilingual classrooms and found the teachers to be proficient in Spanish. Record Vol. 12 at 420, 426-28. His opinion corroborated the testimony of Jacobs. In addition, since 1979, bilingual teachers must have “professional level oral and written proficiency in the language of the target population as measured by an examination approved by the Texas Education Agency.” Castaneda II, TEA Exhibit 1. The record indicates that fifteen out of the twenty-seven RISD bilingual teachers in 1983-84 were hired after TEA implemented this evaluation requirement. Taken together, the above evidence indicates that the RISD bilingual teachers were fluent or proficient in Spanish.
Plaintiffs’ cited evidence fails to undermine the district court’s findings. We have dutifully read the depositions cited by Plaintiffs to support their claim that many teachers admitted their inability to use Spanish. Of the eight depositions mentioned in Plaintiffs’ brief, however, five reveal teachers testifying not that they had language limitations, but instead that they were native speakers of Spanish. A sixth teacher testified that she had learned Spanish as a child and considered herself bilingual. Admittedly enough, two teachers confessed to a lack of proficiency in Spanish. These two teachers, however, no longer teach in the bilingual program at RISD and were not included in the 1983-84 bilingual staff evaluated by the district court.
Of course, fluency in Spanish does not alone constitute an adequate bilingual teacher. Although our directive to the district court to ensure an adequate teacher training program in RISD was premised in part on our perception that the RISD teachers then employed lacked proficiency in Spanish, in-service teacher training procedures are necessary even to native speakers of Spanish to develop and hone the skills necessary to be an adequate bilingual teacher. Since Plaintiffs took the position below that state law is the appropriate model against which to measure RISD’s program and progress, we next consider whether the district court was correct as a matter of law in finding RISD’s training procedures to be in compliance with state law.
*472Under state law, each school district is to provide not fewer than eight days in each school year for in-service training or preparation of teachers. Tex.Educ.Code § 16.052(a) (Vernon Supp.1985). Of these eight days, the regulations provide that at least five are to be devoted to in-service training. After Castaneda I, Superintendent Jacobs journeyed to Austin to consult with experts for suggestions, ideas and support in developing a good in-service training program. Record Vol. 12 at 515. From August 1982 to March 1983, Irma Infante, an instructor in bilingual education courses at Pan American University in Ed-inburg, visited RISD at least five times to lead in-service training sessions. Her main purpose was to improve RISD’s Spanish language instruction in “content” areas, such as social studies, science, and math. A 1982 TEA visit had identified native language “content” instruction as an area requiring improvement. Ms. Infante led question and answer sessions to discover teacher needs, developed units of instruction in various content areas to serve as models and give the teachers a “starting point,” and worked with the teachers, allowing them to place priorities upon the topics about which they wished to learn more. Record Vol. 11 at 351-61. Ms. Infante was accompanied by another instructor at Pan American University, Eva Corano. Record Vol. 11 at 355. We cannot say that the district court erred in holding that these programs satisfy state law requirements. In addition, although Mr. Ma-gallanes had not yet made his final followup visit to RISD at the time of the second trial, he testified that RISD’s progress had been “excellent” and that he expected the program now to be in compliance with state law and comparable to the regular educational program provided non-bilingual children. Record Vol. 12 at 422. Given the above, we affirm the district court in finding RISD’s bilingual program to be in compliance with state law.
V.
Our review on appeal is not to determine whether RISD has taken every possible step to improve the education of its Spanish-speaking students, but only to decide whether the district court’s conclusion that RISD has met the requirements of the fourteenth amendment, Title VI and the EEOA is supported by the record. Most of the district court’s findings bear the imprimatur of Rule 52(a). We are not persuaded that the court’s findings lack record support or that RISD has violated its duty under the fourteenth amendment, Title VI or the EEOA.
For the above reasons, the judgment of the district court is AFFIRMED.
. Our earlier opinion will hereinafter be referred to as Castaneda I. The district court’s unpublished opinion on remand will be referred to as Castaneda II.
. Our review of this case has not been assisted by Plaintiffs’ tendency to cite as support for their arguments passages in the record which often contradict the arguments made.
. Plaintiffs argue that only nine Anglo-American students attend L.C. Smith. Record Vol. 12 at 540-41. However, Castaneda II, Plaintiffs' Exhibit 57, supports the district court’s finding that there are 11 Anglo-American students at L.C. Smith.
. Plaintiffs also allege that "continuing employment discrimination against Mexican-American teachers ... is part and parcel of the historical pattern of discrimination against Mexican-Americans in the RISD.” Appellants’ Brief at 12. We will address the question of RISD’s discrimination in hiring and promotion in Part III of this opinion.
. The Supreme Court reversed and remanded this consolidated case only insofar as the Court of Appeals authorized deferral of student desegregation beyond February 1, 1970.
. Plaintiffs additionally allege discrimination through "classroom grouping" in RISD. It is difficult to understand this claim. Plaintiffs argue that “for the 1983-84 school year, for the kindergarten grades at Pittman School, 91% of the English surnamed children are in classes which are from 40 to 45% English surnamed, even though English surnamed children represent only 17% of the students in the kindergarten class.” Appellants’ Brief at 9. As explained above, however, these kindergarten students are not assigned to “classrooms" per se, but are assigned to “no risk,” "moderate risk” and "high risk" sections on the basis of a battery of nondiscriminatory tests. The 91% English sur-named children cited by Plaintiffs were in the "no risk" ability group sections, Castaneda II, Plaintiffs’ Exhibit 57, and were evenly distributed amongst these sections in various classrooms. The complained-of “classroom grouping” thus becomes a question of ability grouping, which is discussed supra in the text.
. The TEA has divided the state into regional educational service centers. At all times relevant to this case there were twenty educational service centers.
. RISD on appeal urges that the relevant labor market should span the nation. Appellees’ Brief at 30-31. We agree with the district court that the record evidence does not clearly enough define the characteristics of this wide-ranging market.
. 42 U.S.C. § 1983 provides:
Every person who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory or the District of Columbia, subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the constitution and laws, shall be liable to the party implied in an action at law, suit in equity, or other proper proceeding for redress. For the purposes of this section, any Act of Congress applicable exclusively to the District of Columbia shall be considered to be a statute of the District of Columbia.
. The EEOA explicitly provides in § 1703(d) that “discrimination by an educational agency on the basis of race, color, or national origin in the employment ... of faculty or staff" constitutes a denial of equal educational opportunity.
. 42 U.S.C. § 2000d provides:
No person in the United States shall, on the grounds of race, color, or national origin, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any program or activity receiving federal financial assistance.
In Castaneda I, we noted that in order to assert a Title VI claim based on unconstitutional racial discrimination a party must not only allege and prove that the challenged conduct had a differential or disparate impact upon persons of different races, but also assert and prove that the governmental actor, in adopting or employing the challenged practices or undertaking the challenged action, intended to treat similarly situated persons differently on the basis of race. Castaneda I, 648 F.2d at 1000. The Supreme Court has since held that proof of discriminatory intent is not required in a Title VI action for equitable relief. Guardians Association v. Civil Service Commission of the City of New York, 463 U.S. 582, 103 S.Ct. 322, 77 L.Ed.2d 866 (1983). Thus a Title VI action can now be maintained in either the guise of a disparate treatment case, where proof of discriminatory motive is critical, or in the guise of a disparate impact case, involving employment practices that are facially neutral in their treatment of different groups but that in fact fall more harshly on one group than another. In this latter type of case, proof of discriminatory intent is not necessary. See International Brotherhood of Teamsters v. United States, 431 U.S. 324, 335-36 n. 15, 97 S.Ct. 1843, 1854 n. 15, 52 L.Ed.2d 396 (1977) (discussing disparate treatment and disparate impact claims in context of Title VII suits.)
Citing Guardians, the district court observed that it would “address plaintiffs’ claims for equitable relief under Title VI without reference to proof of discriminatory intent.” Record Vol. 1 at 17. The district court’s conclusion that Plaintiffs failed to prove discrimination would seem to encompass the Title VI disparate impact claim. Plaintiffs on appeal briefly urge that their disparate impact claim should have been sustained by the district court. Plaintiffs, however, do not flesh out the elements of this claim. Nor do we think such an exposition would be possible. As we have noted, ”[t]he disparate impact model applies only when an employer has instituted a specific procedure, usually a selection criterion for employment, that can be shown to have a causal connection to a class imbalance in the work force.” Pouncy v. Prudential Insurance Co. of America, 668 F.2d 795, 800 (5th Cir.1982). The plaintiff must show that "a facially neutral employment practice has the result of producing a significantly adverse impact on one race.” Page v. U.S. Industries, Inc., 726 F.2d 1038, 1045 (5th Cir.1984).
Here, Plaintiffs have never alleged the use of a specific procedure or selection criterion neutral on its face. In fact, Plaintiffs have alleged the opposite, arguing that hiring practices and criteria for employment continue to be vague, unorganized and undisclosed with RISD purposely refusing to recruit or hire a sufficient number of Mexican-American teachers. Appellants’ Brief at 25. As Plaintiffs acknowledge, these contentions go to an intentional disparate treatment claim, id., not, we find, to a disparate impact claim. Given the above, we do not find that the district court erroneously applied the law or facts in implicitly denying Plaintiffs relief under a disparate impact theory. The text of this opinion will therefore review only Plaintiffs’ intentional discrimination claims.
. As we have in Part II upheld the district court's conclusion that "RISD has been a unitary system for a sufficient period of time that any vestiges of discrimination have been erased," Record Vol. 1 at 12, RISD does not have a recent history of discrimination such as to warrant a requirement that RISD rebut Plaintiffs' prima facie case by clear and convincing evidence that the challenged employment decisions were motivated by legitimate nondiscriminatory reasons. Castaneda I, 648 F.2d at 994 (citing Lee v. Conecuh County Board of Education, 634 F.2d 959 (5th Cir.1981)).
. At oral argument and in their brief, Plaintiffs discuss an additional school district, Lyford, with a high percentage of Mexican-American teachers. At trial, Superintendent Jacobs briefly noted that Lyford has no movie house and is a small town without significant housing. Record Vol. 12 at 532-33. No representative from the Lyford School District testified. The record does not contain evidence of salary, fringe benefits, or the other factors found to influence a teacher’s employment decision. Ly-ford thus does not prove a helpful comparison to RISD here.
. RISD is not alone in its inability to attract a number of Mexican-Americans from Region I and II schools. Plaintiffs' Exhibit 44 reveals the percentage of Mexican-American teachers employed by RISD and neighboring school districts. • For the 1982-83 school year, RISD employed 39.28% Mexican-American teachers. Sharyland School District employed 42.20%; Rio Hondo, 43.84%; San Perlita, 21.05%. Castaneda II, Plaintiffs Exhibit 44.
. In Castaneda I, we noted that the "record indicates that, as a general rule, the RISD prefers to hire administrative personnel from within the ranks of its current employees.” Because the percentage of Mexican-American administrators was roughly comparable to the percentage of Mexican-American teachers, we concluded that "the statistical evidence in this case would not seem to support an inference of discrimination in promotion, unless, of course, discrimination in hiring is established.” Castaneda I, 648 F.2d at 1003. As we affirm the district court's finding of no discrimination in hiring and as Plaintiffs have not introduced on remand any new evidence of discrimination in pro*470motion, there is no need to address the question of discrimination in promotion here.
. Plaintiffs additionally request the TEA to remedy the "absence of Spanish language curriculum materials" at RISD. The question of curriculum materials, however, was not raised to the district court in Plaintiffs' original pleading, nor mentioned to us in Plaintiffs' brief in Castaneda /. The question of curriculum materials thus played no role in our directives to the district court upon remand and was not there addressed. The issue is therefore not before us now. Even were we to address this issue, Plaintiffs’ evidence is woefully lacking in persuasive power. Plaintiffs apparently introduced no evidence on curriculum materials in the remand trial, and rely here mainly on the testimony of a defense witness, Irma Infante. This testimony does not address the presence or absence of Spanish curriculum materials at RISD, but instead chronicles the efforts of Ms. Infante to help teachers develop their skill at teaching "content” area subjects in Spanish. See text infra.