KING, Circuit Judge:
The plaintiffs1 in this action allege that the Austin Independent School District (AISD), by instituting a new student assignment plan, violated the Equal Protection Clause of the U.S. Constitution by returning to the former dual school system which had been held unitary in 1983. After a two-day trial, the district court entered judgment for AISD.2 Price v. Austin In-dep. School Dish, 729 F.Supp. 533 (W.D.Tex.1990). In a thorough opinion, the court held that the plaintiffs in this case “failed to establish that ... [AISD] acted with the intent to discriminate against racial or ethnic minority persons in violation of the Fourteenth Amendment to the United States Constitution.” Id. at 552-53.
The plaintiffs appeal the judgment, claiming that the district court failed to shift the burden of proof to AISD once the plaintiffs had demonstrated a prima facie case of discrimination. The plaintiffs also claim that the district court, in finding that AISD had acted free of discriminatory intent, relied on improper evidence, failed to consider AISD’s history of purposeful discrimination, and gave too much weight to the earlier finding of unitariness.
We find that the district court’s legal analysis is consistent with the law of the United States and this circuit. We also hold that the district court’s findings of fact, based on the testimony presented and the record considered as a whole, were not clearly erroneous, and we therefore AFFIRM the judgment below.
I. BACKGROUND
A. The Previous Litigation
In 1970, the United States, acting under the authority granted to the Attorney General by the Civil Rights Act of 1964, 42 U.S.C. § 2000c-6(a) and (b), brought suit against the Texas Education Agency and AISD. The suit alleged that AISD operated a dual school system in violation of the U.S. Constitution. In 1971, this court granted a motion permitting the intervention of a group of black and Mexiean-American parents. See United States v. Texas Educ. Agency, 467 F.2d 848, 853 n. 1 (5th Cir.1972) (Austin I). During the 1970s, this court repeatedly held that AISD had intentionally discriminated against black and Mexican-American students.3
In November 1979, the district court entered a memorandum opinion and order calling for the submission of an integrative student assignment plan. On January 2, 1980, the district court approved a consent decree, agreed to by all the parties in the litigation.
[1311]*1311The 1980 consent decree represented the resolute commitment of all parties to transform AISD into a unitary school system. The decree provided for the implementation of busing, alteration of attendance-zone lines, as well as other integrative actions. Under the decree, the [district court] would retain jurisdiction over the case for three years. Upon the expiration of the three-year period, subsequent to notice and the opportunity to object, AISD would be declared a unitary school system and the case dismissed.
Price, 729 F.Supp. at 535.4
The plaintiff-intervenors in 1983 objected to the scheduled declaration of unitariness, but later withdrew their objections and all the parties filed an agreed motion to dismiss, accompanied by a stipulation. This decree declared AISD unitary and dismissed the case. The stipulation, which continued in effect until September 1986, granted the plaintiff-intervenors the right to a hearing if AISD changed the student assignment plan in a way which discriminated against students on the basis of race, color or national origin. At such a hearing, AISD would be required to show cause why the case should not be reopened. In September 1986, by its own terms, the stipulation’s period of effect ended, and it was no longer enforceable. Id.; see also United States v. Overton, 834 F.2d 1171, 1174 (5th Cir.1987).
B. The 1987 Student Assignment Plan
In April 1987, the Board adopted a revised student assignment plan, to become effective in the 1987-88 school year. The new plan primarily affected elementary schools, and eliminated crosstown busing (which had been a part of the 1980 consent decree, and was designed to further desegregation) of most students in pre-kinder-garten through fifth grade.
The plan would result in sixteen elementary schools which would have predominantly minority student populations (90% or more, according to the district court’s opinion). See 729 F.Supp. at 539 n. 9.5 These sixteen schools were to be the subject of additional funding and other preferential treatment, such as a choice of teachers and staff. This preferential treatment was known as the “Plan for Educational Excellence,” or the “Priority Schools Program.”
The 1987 plan also included other elements. It retained the district’s policy of majority-minority transfer, by which “any student whose race or ethnicity constitutes the majority of the school’s population [may] transfer with free transportation to a school where the student’s racial or ethnic group is a minority of the school’s population.” Id. at 538. Furthermore, the plan permitted any student, who had been bused for integration purposes under the 1980 consent decree, to remain at the school to which he or she had been assigned under that decree.6
The 1987 student assignment plan was the result of two related concerns. First, the need for revised student assignment boundary changes had been a subject of discussion since 1985. In June 1986, the Board adopted a set of nine criteria to be [1312]*1312used by AISD staff in formulating student assignment and boundary proposals. See id. at 537. These criteria included a target of ethnicity in each school of within ten percentage points of the districtwide proportions. The criteria also sought to limit the extent and length of busing as a means of achieving the necessary integration. One of these criteria stated that “[i]f students must be bused in order to attain the target at each school, grades 6 through 12 should be bused[.]” Id. The criteria also provided for the adoption of the middle school structure, a change implemented by the 1987 plan. Finally, the criteria called for “additional resources to provide an enriched learning environment” at schools which did not meet the target for ethnicity and which had “high concentrations of low achieving/low-income students[.]” Thus, the 1987 plan substantially followed the criteria adopted by the Board. The district court found that the second factor leading to the adoption of the 1987 plan was “a conscious decision by the Board to establish neighborhood elementary schools in AISD.” Id. at 538.
C. The Present Suit
In 1987, the plaintiffs-intervenors in Austin I-IV sought to reopen that case in order to challenge the new student assignment plan. The district court dismissed that attempt, holding that anyone seeking to challenge the new student assignment plan must file a new lawsuit. The plaintiffs in this case then took the field by filing a new complaint seeking, inter alia, a preliminary injunction against the implementation of the new student assignment plan. The district court denied the injunction. In a consolidated appeal, we affirmed both the refusal to reopen the earlier litigation and the denial of a preliminary injunction. United States v. Overton, 834 F.2d 1171, 1177-78 (5th Cir.1987), aff'g 671 F.Supp. 484 (W.D.Tex.1987).
After the case returned to the district court, nearly two years of discovery and pre-trial proceedings ensued. The case was tried before the district court November 6-8, 1989. The plaintiffs and defendants produced a total of twenty witnesses, and the court received a significant number of exhibits into the record. After evaluating the testimony, the district court entered judgment for AISD. The district court’s opinion thoroughly outlined the evidence presented, and stated the court’s reasons for attributing relative credibility to various witnesses. The court concluded that the Board had not adopted the 1987 plan with the intent to discriminate against blacks or Mexican-Americans, and that valid educational concerns were served by the plan’s adoption and implementation. Price, 729 F.Supp. at 549.
II. STANDARD OF REVIEW
We review the district court’s factual findings for clear error. The determination that AISD did not purposefully discriminate in violation of the Constitution is a factual finding, reviewable only under the clearly erroneous standard. Anderson v. Bessemer City, 470 U.S. 564, 573, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985); Pullman-Standard v. Swint, 456 U.S. 273, 287, 102 S.Ct. 1781, 1789, 72 L.Ed.2d 66 (1982); Norris v. Hartmarx Specialty Stores, Inc., 913 F.2d 253, 255 (5th Cir.1990). “If the district court’s findings are plausible in light of the record viewed in its entirety, we must accept them, even though we might have weighed the evidence differently if we had been sitting as a trier of fact.” Norris, 913 F.2d at 255.
Rule 52(a) emphasizes the deference to be given the trial judge's findings when they rest on credibility determinations: “due regard shall be given the opportunity of the trial court to judge the credibility of the witnesses.” Fed.R.Civ.P. 52(a). We therefore must apply the clear error standard with particular care in cases involving demeanor testimony. See, e.g., Anderson, 470 U.S. at 575, 105 S.Ct. at 1512; Verrett v. McDonough Marine Serv., 705 F.2d 1437, 1443 (5th Cir.1983). Of course, we review the district court’s legal conclusions de novo.
III. ANALYSIS
The plaintiffs raise four arguments on appeal. First, as a legal matter, they ar[1313]*1313gue that findings made in the previous litigation, that AISD had intentionally discriminated against black and Mexican-American students, are binding on the defendants and demonstrate a history of de jure segregation by AISD. The plaintiffs seek to invoke this history in order to shift the burden of proof to AISD upon proof of discriminatory effect. On the plaintiffs’ theory, AISD would be required to demonstrate its absence of discriminatory intent in adopting the new student assignment plan. Second, as a factual matter, the plaintiffs assert that AISD’s history of purposeful discrimination (as found in the previous litigation) was not given sufficient weight by the district court. Relatedly, they argue that the district court gave too much weight to the finding of unitariness entered in the previous litigation.
Third, they contend that the district court erred in considering subjective evidence of the Board members’ good faith. They assert that only circumstantial evidence of discriminatory intent should have been considered. Finally, they argue that the district court failed to give proper weight to the evidence presented. They point to a number of alleged discrepancies between the district court’s findings of fact and its conclusion that AISD acted without discriminatory intent.
A. History and the Burden of Proof
1. Burden shifting: the plaintiffs’ arguments
The plaintiffs argue that the findings of intentional discrimination made against AISD in the prior suit collaterally estop AISD from denying that it previously operated a dual school system in violation of the Constitution. They rely on the judicially found fact of historical segregation as the first step in an argument which seeks to invoke a burden-shifting analysis applicable to pre-unitary school districts. They argue that since AISD has been found to have operated a dual school system, it remains under a continuing obligation to eradicate the vestiges of its past discriminatory actions and is proscribed from taking even facially neutral actions which serve to reestablish the dual school system. Thus, once the plaintiffs have demonstrated a prima facie case of discriminatory effect, they assert, the burden of proof should have shifted to AISD to explain why its decision was not intentionally discriminatory.
This burden-shifting analysis applies in cases of dual school systems prior to a finding of unitariness. See Dayton Board of Educ. v. Brinkman, 443 U.S. 526, 537, 99 S.Ct. 2971, 2979, 61 L.Ed.2d 720 (1979) (Dayton II); Keyes v. School Dist., 413 U.S. 189, 208-09, 93 S.Ct. 2686, 2697-98, 37 L.Ed.2d 548 (1973); see also Swann v. Charlotte-Mecklenburg Board of Educ., 402 U.S. 1, 18, 91 S.Ct. 1267, 1277, 28 L.Ed.2d 554 (1971). It is clear, however, that the analysis does not apply once a finding of unitariness has been entered. In a later suit, the plaintiffs bear the burden of proving that the school board acted with the intent to discriminate. Overton, 834 F.2d at 1174-77; Riddick v. School Board, 784 F.2d 521, 534 (4th Cir.), cert. denied, 479 U.S. 938, 107 S.Ct. 420, 93 L.Ed.2d 370 (1986). The district court applied this burden of proof properly; indeed, it found that this was “the sole question presented to [it] for resolution in this case_” Price, 729 F.Supp. at 536.
The plaintiffs urge this court to hold that, notwithstanding the finding of unitariness, AISD’s adoption of the new student assignment plan violated its duty to eliminate all vestiges of past discrimination and to refrain from taking actions which perpetuate the effects of past racial segregation. We do not reach the issue of whether the 1987 student assignment plan would have breached such a duty. It is clear that AISD was under no such obligation after 1983, when the district court found that it had achieved unitary status and dismissed the case.
In urging the contrary, the plaintiffs rely in their briefs on Dowell v. Board of Educ., 795 F.2d 1516 (10th Cir.), cert. denied, 479 U.S. 938, 107 S.Ct. 420, 93 L.Ed.2d 370 (1986), and later proceeding, 890 F.2d 1483 (10th Cir.1989), rev’d, — [1314]*1314U.S. -, 111 S.Ct. 630, 112 L.Ed.2d 715 (1991). Dowell held that a declaration of unitariness does not foreclose the district court’s continued supervision of a school district. The U.S. Supreme Court, during the pendency of this appeal, reversed the Tenth Circuit on the issue of when an injunction ends.
In its opinion, the Supreme Court acknowledged that the Fifth Circuit uses the term “unitary” “to identify a school district that has completely remedied all vestiges of past discrimination.” 111 S.Ct. at 635 (citing Overton, 834 F.2d at 1175). The Court also noted that the law of the Fourth Circuit is in accord with ours. Id. (citing Riddick, 784 F.2d at 533-34). The Court contrasted this usage with the meaning attributed to “unitary” by the Tenth Circuit. The Tenth Circuit “ha[s] used ‘unitary’ to describe any school district that has currently desegregated student assignments, whether or not that status is solely the result of a court-imposed desegregation plan. In other words, such a school district could be called unitary and nevertheless still contain vestiges of past discrimination.” Id. (citation omitted). The Court explicitly declined to rule on the “correct” usage of the term, and merely noted the discrepancy. In the absence of a definitive pronouncement to the contrary by the Supreme Court, we are bound by our precedent, and cannot accept the reasoning of cases decided by circuit courts whose jurisprudence diverges from our own.
2. Unitariness: the law of the Fifth Circuit
The Court’s statement of the law in the Fifth Circuit is clearly correct. We have stated that unitariness represents a finding “that the school district has eliminated all aspects of de jure segregation _” Ross v. Houston Indep. School Dist., 699 F.2d 218, 219 (5th Cir.1983). Of course, this will not always mean a state of complete integration. Ross upheld a district court’s determination that Houston Independent School District was unitary, in spite of the fact that the student populations of nearly ten percent of the schools in the system had been 90% or more black since 1960. Id. at 226. We stated that the appropriate measure of unitariness was “whether the past has been eradicated so far as it remains in the power of school officials and courts to do so[.]” Id. at 227.
This “practicability” test permits a finding of unitariness where the school district has done all that it could to remedy the segregation caused by official action. It recognizes that in some unique circumstances, some segregation will remain; nonetheless, “immutable geographic factors and post-desegregation demographic changes that prevent the homogenation of all student bodies do not bar judicial recognition that the school system is unitary.” Id. at 225. The district court here found that “the location of concentrations of racially or ethnically identifiable residents in particular geographical areas of the Austin metropolitan area together with changes and shifts in these population concentrations are common to most cities and result from economic conditions and cultural interests outside the control of AISD.” 729 F.Supp. at 550. The Ross practicability qualifier was implicitly approved by the Supreme Court in its recent reversal of the Tenth Circuit, when it stated that the district court there should “eonsider[ ] whether the vestiges of de jure segregation had been eliminated as far as practicable.... ” Dowell, 111 S.Ct. at 638.
3. Effect of unitariness: the circuit split
We have also held that a finding of uni-tariness “means that a school board is free to act without federal supervision so long as the board does not purposefully discriminate; only intentional discrimination violates the Constitution.” Overton, 834 F.2d at 1175. This holding is consonant with our earlier jurisprudence. In one case, we had stated that a unitariness finding “is critical because once it is made a federal court loses its power to remedy the lingering vestiges of past discrimination absent a showing that either the school authorities or the state had deliberately attempted to fix or alter demographic patterns to affect [1315]*1315the racial composition of schools.” United States v. South Park Indep. School Dist., 566 F.2d 1221, 1224 (5th Cir.), cert. denied, 439 U.S. 1007, 99 S.Ct. 622, 58 L.Ed.2d 684 (1978).
The Tenth Circuit disagrees with us regarding the effect of a finding of unitariness. Dowell held that the burden-shifting analysis of Keyes continued to apply to a formerly dual school district, even after a finding of unitariness and dismissal of the case. Dowell, 795 F.2d at 1523. We note at the outset that this holding may be affected by the Supreme Court's reversal of the Tenth Circuit on the related issue of the continued viability of an injunction. See Dowell, 111 S.Ct. at 637. We also emphasize that the Tenth Circuit’s jurisprudence in this area differs sharply from our own, as well as from that of the First and Fourth Circuits. Cf. Morgan v. Nucci, 831 F.2d 313, 318 (1st Cir.1987); Riddick, 784 F.2d at 532-34. This circuit split has been noted by courts and commentators faced with the problem. See, e.g., United States v. Lawrence County School Dist., 799 F.2d 1031 (5th Cir.1986); Lee v. Macon County Board of Educ., 681 F.Supp. 730, 737 (N.D.Ala.1988); Annotation, Circumstances Warranting Judicial Determination or Declaration of Unitary Status with Regard to Schools Operating Under Court-Ordered or — Supervised Desegregation Plans and the Effects of Such Declarations, 94 A.L.R.Fed. 667, 723-25 (1987); Note, Allocating the Burden of Proof After a Finding of Unitariness in School Desegregation Litigation, 100 Harv.L.Rev. 653 (1987). As noted above, however, the law of the Fifth Circuit is clear on this matter, and we are bound to follow it in the absence of Supreme Court pronouncements to the contrary.
In this instance, we are bound additionally by the law of the case. Our holding in Overton, that “only intentional discrimination violates the Constitution,” 834 F.2d at 1175, was a “decision of a legal issue by an appellate court establishing] the ‘law of the case’ and must be followed in all subsequent proceedings in the same case at both the trial and appellate levels.... ” Schex-nider v. McDermott Int’l, Inc., 868 F.2d 717, 718 (5th Cir.), cert. denied, 493 U.S. 851, 110 S.Ct. 150, 107 L.Ed.2d 108 (1989); see also Pegues v. Morehouse Parish School Board, 706 F.2d 735, 738 (5th Cir.1983); White v. Murtha, 377 F.2d 428, 431-32 (5th Cir.1967). Overton’s pronouncement binds our consideration of this issue on this subsequent appeal in the same case.
4. Scope of the unitariness finding
The finding of unitariness by the district court in 1983 represented a historical snapshot. It was a finding made at a single point in time, and obviously does not lend any extraordinary presumption of legality to later Board actions. In fact, it grants the Board no presumption at all; it merely removes the federal courts from their supervisory role over the Board’s decisions.
We laid out quite clearly in Overton the test for determining the constitutionality of Board actions taken following the finding of unitariness. 834 F.2d at 1178. The requirement that the plaintiffs here prove “intentional discrimination,” Id. at 1175, does not impose on them an insurmountable burden. Indeed, this is the standard test for alleged violations of the Equal Protection Clause. See Arlington Heights v. Metropolitan Housing Dev. Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977); Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976).
Of course, the finding of unitariness applies only to AISD’s actions taken up to the time of the 1983 consent decree. The finding does not handicap the plaintiffs in this action with respect to their primary obligation, the need to show that AISD acted with the intent to discriminate in adopting the 1987 student assignment plan. We merely hold, as the district court did, that plaintiffs must prove intentional discrimination without relying on the crutch of a presumption imported from a previous action.
B. Unitariness and the Role of History
The plaintiffs argue further that, as a factual matter, the district court failed to [1316]*1316give adequate weight to the findings of segregation and intentional discrimination made in the prior suit. This amounts to an assertion that the previous findings, when seen in the light of the discriminatory effect of the new student assignment plan, demonstrate the requisite purposeful discrimination. They assert that the 1979 findings of AISD’s complicity in fostering residential segregation7 necessitate a finding that the segregative effects of the 1987 student assignment plan were directly caused by the pre-1980 Board’s segre-gative actions. They allege that such a finding would itself require a concomitant finding of purposeful discrimination with regard to the 1987 plan.
The plaintiffs cannot gainsay that the presence or absence of discriminatory purpose is a factual issue to be determined by the district court on the basis of all the available evidence. Anderson, 470 U.S. at 573, 105 S.Ct. at 1511; Pullman-Standard, 456 U.S. at 287, 102 S.Ct. at 1789. We agree, as the district court did, that the history of segregation by AISD is a relevant factor to be considered in determining the presence or absence of purposeful discrimination with regard to the 1987 plan. See Arlington Heights, 429 U.S. at 267-68, 97 S.Ct. at 564-65. However, the plaintiffs cannot avoid their burden of proof in this case merely by pointing to findings entered in the prior litigation. We have already demonstrated that the 1983 finding of unitariness removed any presumption of illegality from Board actions, even if they had a discriminatory effect. Thus, the 1979 findings of segregation by AISD, both in the schools and in residential neighborhoods, constitute appropriate historical factors to be considered by the district court, but do not, in themselves, end the inquiry.
The district court, in considering the problem of residential segregation, stated that “immutable geographic factors which prevent the homogenation of all student bodies do[] not preclude the perpetuation of unitary status.” Price, 729 F.Supp. at 550 (citing Ross, 699 F.2d at 225). The court found that “the location of concentrations of racially or ethnically identifiable residents in particular geographical areas of the Austin metropolitan area together with changes and shifts in these population concentrations are common to most cities and result from economic conditions and cultural interests outside the control of AISD.” Id.; see also Pasadena City Board of Educ. v. Spangler, 427 U.S. 424, 435-36, 96 S.Ct. 2697, 2704-05, 49 L.Ed.2d 599 (1976) (quoting Swann, 402 U.S. at 31-32, 91 S.Ct. at 1283-84).
The plaintiffs contend that this conclusion is not supported by the evidence presented at the trial. On cross-examination, however, the plaintiffs’ expert, Yale Rabin, acknowledged that, in some respects at least, the 1980 census indicated a greater dispersion of blacks into areas beyond East Austin.8 Moreover, the census data on which Professor Rabin relied was nearly ten years old by the time of the trial. See Price, 729 F.Supp. at 540. The district court’s findings, supported by Professor Rabin’s testimony, demonstrate an “encouraging but slow ongoing dispersion of Black persons outside areas of previous concentration into areas formerly dominated by majority persons.” Id. Professor Rabin’s direct examination testimony also reflects a significant increase in black population in an area which had previously been “a moderate income white residential [1317]*1317area.” Clearly, this supports the district court’s finding that the “changes and shifts in these population concentrations ... resulted] from economic conditions and cultural interests outside the control of AISD.” Id. at 550.
In short, the district court appropriately recognized “the segregative history of [AISD] prior to 1980,” as well as “its unitary status by virtue of the Consent Decree of 1983.” Id. at 536. The court explicitly acknowledged the validity of the 1979 findings of prior discrimination. Id. at 540. During the trial, the judge specifically stated that “as far as relevancy is concerned[,] ... the court would accept the [1979] findings of this court ... with regard to [1979 conditions in the school district].”
It is clear to us that the district court considered the evidence of prior discrimination as a relevant factor in the determination of whether AISD acted with discriminatory purpose. It is equally clear that history, like discriminatory effect, is not dispositive of the question of intent. Both are factors to be weighed in the balance by the district court; neither is in itself proof of the forbidden discriminatory intent. See Arlington Heights, 429 U.S. at 266-68, 97 S.Ct. at 564-65.
Based on an independent review of the record as a whole, we cannot conclude that the weight given by the district court to the factors of historical discrimination and unitary status resulted in findings that were clearly erroneous. The plaintiffs point to no clear error in the district court opinion; they merely seek to have this court reassess the credibility of witnesses and the overall weight of the evidence. This we cannot do. Where the evidence below supports two alternative interpretations, the appellate court may not substitute its judgment for the district court’s. See Anderson, 470 U.S. at 573-75, 105 S.Ct. at 1511-12; Norris, 913 F.2d at 255; Fed.R.Civ.P. 52(a).
C. Proper Evidence of Intent
The plaintiffs contend that the district court erred in considering subjective evidence of the Board’s good faith intent, including testimony of individual Board members. They assert that discriminatory intent should be measured by “institutional intent based on circumstantial evidence, historical and otherwise.” To a degree, this smacks of the Dowell presumption based on pre-unitariness history of segregation. We have already noted that this argument will not avail the plaintiffs here.
Nonetheless, to the extent that the plaintiffs contend, as a factual matter, that the district court should have considered only circumstantial evidence of discriminatory intent, we consider this argument separately. The plaintiffs do not suggest that the testimony of individual Board members was inadmissible; indeed, an important witness on behalf of the plaintiffs was Abel Ruiz, himself a former Board member. Rather, they argue that the district court gave too much weight to the subjective motivations of individual Board members.
The plaintiffs provide quotes from a number of cases which seem to suggest that the presence or absence of discriminatory intent must be determined solely by reference to circumstantial evidence. For the most part, these quotes lack necessary context, although a few do address the point urged by the plaintiffs. One passage, from the Tenth Circuit’s first opinion in Dowell, criticizes the district court in that case for basing its finding of no intentional discrimination on testimony elicited by asking “various witnesses, lay and expert, if they believed the Plan was adopted with discriminatory intent.” 890 F.2d at 1502. The court stated that “the quest for discriminatory intent is not so straightforward.” Id. Instead, the court suggests, the district court should itself analyze all of the available evidence, and determine the presence or absence of discriminatory intent.
The flaw in the Dowell district court’s analysis was-, that it shirked the burden of determining, based on the evidence presented, whether the plan had been adopted with discriminatory intent. Instead, it relied on the opinion testimony of various witnesses that no such intent was present. Here, the district court analyzed all of the testimony [1318]*1318presented, and did not rely on opinion testimony of lay or expert witnesses as to the presence or absence of discriminatory intent. The plaintiffs here do not assert that the district court made this type of error. The court’s opinion that the Board acted without discriminatory intent was based not on bald assertions to that effect by Board members or by others, but on the credibility of the testimony elicited from many witnesses that the new student assignment plan had been adopted in order to fulfill certain legitimate educational purposes.
The Tenth Circuit, in Dowell, and the plaintiffs also cite a district court proceeding where the district court held that the relevant intent could not be established by subjective good faith evidence, but must be shown only by circumstantial evidence. Keyes v. School Dist., 670 F.Supp. 1513, 1516 (D.Colo.1987), aff'd, in part and remanded, 895 F.2d 659 (10th Cir.1990), cert. denied, — U.S. -, 111 S.Ct. 951, 112 L.Ed.2d 1040 (1991), quoted in Dowell, 890 F.2d at 1503. Keyes, however, involved a pre-unitary school district, in which the presumption of discriminatory intent attaches on a showing of prima facie segregative effect. We have already held that, following a finding of unitariness and dismissal of the case, this presumption no longer applies. The methodology of Keyes is inapplicable to this case.
A Ninth Circuit case, reversing a district court’s finding of no discriminatory intent, stated that “[ojrdinarily, only circumstantial evidence is available to establish segregative intent.” Diaz v. San Jose Unified School Dist., 733 F.2d 660, 662 (9th Cir.1984), cert. denied, 471 U.S. 1065, 105 S.Ct. 2140, 85 L.Ed.2d 497 (1985). However, the sentence states a descriptive, not a normative, proposition. The posture of the case suggests that the court was justifying its reversal of the district court, in spite of the absence of any direct evidence of discriminatory intent, by pointing out the rarity of any such evidence. The school district in Diaz was under an affirmative duty, imposed by state law, to alleviate school segregation, regardless of its cause. Id. at 665-66. Here, the finding of unitariness removed AISD’s affirmative duty which had previously attached upon a finding of liability in the prior suit. While the empirical validity of the Ninth Circuit’s statement may be largely undisputed, since smoking gun evidence in desegregation cases is rare, the statement simply does not assist the decisional process in this case.
The plaintiffs cite Arlington Heights for the proposition that historical background and the sequence of events leading to a decision are important eviden-tiary points in seeking to determine the presence or absence of discriminatory intent. 429 U.S. at 267-68, 97 S.Ct. at 564-65. In the present case, of course, any pre-1983 historical evidence must be tempered by the finding of unitariness. Thus, the plaintiffs here must rely on post-1983 circumstantial evidence to prove the discriminatory intent of AISD in adopting the new student assignment plan. The bulk of the evidence relied on by plaintiffs, however, merely demonstrates the segregative impact of the new plan. The Supreme Court stated in Arlington Heights that “impact alone is not determinative, and the Court must look to other evidence.” Id. at 266, 97 S.Ct. at 564 (footnote omitted). The reference to the various forms of circumstantial evidence in Arlington Heights was necessary, the court found, because the testimony of the actual decisionmakers “frequently will be barred by privilege." Id. at 268, 97 S.Ct. at 565. Here, however, the Board members testified fully without invoking any privilege. To the extent that the justifications advanced in their testimony do not demonstrate a pretext for intentionally discriminatory actions, the logic of Arlington Heights suggests that the evidence relied on by the district court is actually stronger than the circumstantial evidence proffered by the plaintiffs. When the district court looked at all the evidence in this case, it found that AISD had not acted with discriminatory intent. Plaintiffs have not demonstrated that, on the evidence presented, this finding is clearly erroneous.
[1319]*1319The plaintiffs also cite Personnel Administrator v. Feeney, 442 U.S. 256, 279 n. 25, 99 S.Ct. 2282, 2296 n. 25, 60 L.Ed.2d 870 (1979), asserting that reasonable foreseeability of discriminatory impact leads to a strong inference that the adverse effects were intended. This argument falls of its own weight, however, since Feeney stands directly for the proposition that foreseeability of discriminatory impact, without more, does not constitute the forbidden discriminatory purpose. “ ‘Discriminatory purpose,’ however, implies more than intent as volition or intent as awareness of consequences. It implies that the decisionmaker ... selected or reaffirmed a particular course of action at least in part ‘because of,’ not merely ‘in spite of,’ its adverse effects upon an identifiable group.” Id. at 279, 99 S.Ct. at 2296. The district court properly cited Feeney for this proposition. Price, 729 F.Supp. at 549.
Indeed, the sentence quoted by the plaintiffs, that inevitable adverse consequences suggest a strong inference of intent, is immediately followed by a caveat:
But in this inquiry — made as it is under the Constitution — an inference is a working tool, not a synonym for proof. When, as here, the impact is essentially an unavoidable consequence of a legislative policy that has in itself always been deemed to be legitimate, and when, as here, the statutory history and all of the available evidence affirmatively demonstrate the opposite, the inference simply fails to ripen into proof.
Feeney, 442 U.S. at 279 n. 25, 99 S.Ct. at 2296 n. 25. The court here found that the policies of minimizing transportation time for elementary students, and of instituting neighborhood schools for those students, were legitimate ones. The court’s opinion details the evidence relied on to demonstrate “the opposite” of what plaintiffs contend.
D. Weight of the Evidence
The plaintiffs point to four areas in which they allege that the district court improperly weighed the evidence presented. First, they emphasize that the new student assignment plan had the effect of returning thirteen schools in East Austin which had been racially identifiable before the 1980 consent decree to that status again. Second, they contend that the Board failed to consider alternative plans which would have achieved the same educational purposes as the new student assignment plan, but with less segregative effect. Third, they note that the district court credited evidence that racially identifiable schools were disproportionately in need of repair or replacement. They contend that the district court should have found discriminatory intent on the basis of this testimony. Finally, they point to a discrepancy between the district court’s opinion and the evidence presented.
1. Similarity between 1987 and 1979
The plaintiffs point to the similarities between the 1987 student assignment plan and the pre-1980 plan held unconstitutional in the prior litigation. They contend that the same schools were racially identifiable under both plans, with the only exception being that the number of racially identifiable schools had increased to accommodate the larger minority population. Consequently, they argue, the inference of intentional discrimination should be stronger since the new plan directly parallels the unconstitutional one.
The plaintiffs assert, without citation, that “[i]t was plain error for the court to ignore the stark historical connection between the segregation of 1979 and the segregation of 1987.” Hyperbole aside, the plaintiffs do not point to any authority for their contention that such a connection must necessarily lead to a finding of intentional discrimination. The historical connection asserted by the plaintiffs, according to Arlington Heights, constitutes “one evidentiary source” among many. 429 U.S. at 267, 97 S.Ct. at 564.
Indeed, in order to argue this point, the plaintiffs must again rely on the presumption of discriminatory intent which attaches to a pre-unitary segregated school district. We have already held that such an analysis, employed by the Tenth Circuit in Dow-ell, is not the law of this circuit. Overton, [1320]*1320834 F.2d at 1175-77. As we have described above and previously held, the burden of proof in this ease is on the plaintiffs to demonstrate intentional discrimination, and they may not rely on presumptions applicable to pre-unitary school districts.
The district court duly acknowledged the history of segregation by AISD generally. Price, 729 F.Supp. at 540. The court also correctly held that “[t]he increase in number of racially identifiable schools under the 1987 plan, without more, amounts to evidence of disparate impact but not intentional discrimination.” Id. at 550. Although the district court did not make the historical connection urged by the plaintiffs, we cannot say that it was clear error not to do so.
2. Consideration of alternative plans
The plaintiffs allege that the Board failed to consider alternative plans which would accomplish the same educational goals but result in less segregative effect. However, this assertion is directly contradicted by the testimony of Dan Robertson, one of the two AISD staff members primarily responsible for drawing up the 1987 plan. Mr. Robertson testified that a number of alternative plans were devised and considered by the AISD staff, and ultimately were presented to the. Board. These plans consisted of various elementary school boundary alternatives and many of the plans were introduced at the trial in this case as exhibit evidence. Mr. Robertson’s testimony asserts that the planners “attempt[ed] to maximize integration within those criteria [adopted by the Board].”
The plaintiffs rely on the courtroom testimony of Ed Small, a former Board member and president, as well as the deposition testimony of John Lay, also a former Board member. Mr. Small testified that the Board did not ask specifically for additional alternative plans which would have the effect of reducing the number of racially identifiable schools. However, Mr. Small’s testimony was given in response to a question which asks about alternatives to a specific criterion (that, if busing is necessary, it should involve students in grades six through twelve). The testimony does not refer to alternative global plans, whereas Mr. Robertson’s testimony explicitly does. Mr. Small also testified that the Board believed that the 1986 criteria themselves included a requirement of integration, to the extent possible in conformity with the remaining criteria.
Mr. Lay similarly testified, at his deposition, that the Board did not request AISD staff to provide alternative plans which would have resulted in some busing of elementary school students. However, his testimony is consistent with the criteria adopted by the Board in 1986, which required that, if busing were necessary to achieve the desired racial and ethnic proportions, only students in grades six through twelve should be bused. This criterion, as the district court found, is consistent with the approach used by the Board and by the AISD staff. That approach continued to use busing for integration purposes in grades six through twelve (as well as in a few situations in the lower grades), but attempted to reconfigure the school boundaries for elementary schools so as to come as close as possible to the ethnicity targets stated in the criteria. See Price, 729 F.Supp. at 544. The plaintiffs do not argue that the criteria themselves were adopted with discriminatory purpose or intent. The absence of alternative plans which exceeded the bounds of the criteria explicitly adopted for these purposes, then, cannot be said to indicate purposeful discrimination.
The exhibits admitted include a number of variations on the boundary lines for various schools. They do not include any corresponding indication of the racial proportions which would result from such boundaries, however. Consequently, it is impossible to tell, based on the record before this court, whether the student assignment plan actually adopted by the Board was more or less segregative than the alternatives presented to the Board by AISD staff. Given the testimony presented and the exhibits admitted, however, the plaintiffs’ claim that the Board failed to consid[1321]*1321er less segregative alternatives is insupportable.
3. Disrepair of minority schools
The district court explicitly credited the testimony of Laurie Jones, a former elementary teacher and principal in the East Austin area. Ms. Jones testified that “many predominantly minority student elementary schools in Austin are in less than adequate repair and are in immediate need of repair, modernization or construction.” Price, 729 F.Supp. at 543. The court also found “a continued lack of attention by [AISD] to the improvement and physical maintenance of elementary school buildings and facilities located in or bordering on the inner city areas of Austin.” Id. at 547. The court found, however, that “this evidence does not at this time clearly demonstrate an intent to discriminate against minority students at those grade levels,” and attributed the cause of the disrepair to “AISD’s lack of adequate financial re-sources_” Id. at 547-48.
The plaintiffs argue that the court “inexplicably failed to relate this condition [of disrepair] to the judicially established history of discrimination or to otherwise give this factor its proper weight in determining institutional discriminatory intent.” In its first aspect, this argument again seeks merely to invoke the pre-unitariness presumptions and burden-shifting which we have explicitly disavowed. The broader concern that this factor, since credited by the court, should have been determinative of a finding of discriminatory intent, cannot stand. The court’s finding of no intentional discrimination was made after review of a broad spectrum of evidence concerning multiple factors. The plaintiffs’ assertion that discriminatory impact in school repair and replacement constitutes intentional discrimination can no more be true than any other showing of discriminatory impact, which the Supreme Court has held generally insufficient for a showing of intentional discrimination.9 See, e.g., Feeney, 442 U.S. at 274-79, 99 S.Ct. at 2293-96; Arlington Heights, 429 U.S. at 266, 97 S.Ct. at 564; Swann, 402 U.S. at 26, 91 S.Ct. at 1281.
4. Abel Ruiz’s testimony
The plaintiffs point to a discrepancy between the district court’s factual findings and the evidence in the record. The district court generally discredited the testimony of Abel Ruiz, a former Board member. One of the reasons given for doing so was Mr. Ruiz’s testimony regarding the criteria for developing a new plan; the criteria were adopted by the Board on Mr. Ruiz’s motion in 1986.
Mr. Ruiz testified that he introduced the criteria with the intention that criterion number one (requiring ethnic balance in each school within ten percentage points of the district’s as a whole) should override criterion number six (that if busing is required for integration purposes, it should involve students in grades six through twelve only). The district court stated that “[tjhere are no statements apparent in Board Minutes admitted into evidence that indicate Ruiz ever vocalized his subjective belief as to the priority that should be given each criteria [sic]_” Price, 729 F.Supp. at 542-43.
The plaintiffs point out that the minutes do reflect, albeit quite ambiguously, Mr. Ruiz’s intent as to priority of the criteria.10 Even if this were a matter where the discrepancy were clearer, we do not think the error significant in terms of the determina[1322]*1322tion of the presence or absence of discriminatory intent. The district court found Mr. Ruiz’s testimony unpersuasive for numerous reasons, which the opinion carefully details. See id. at 541-43. íhe lack of credibility attributed to Mr. Ruiz was, in turn, only one of many factors which led to the district court’s finding of no discriminatory intent.
IV. CONCLUSION
The district court’s application of the law in this case was proper. On our consideration of the entire record, we find that the district court’s factual conclusions were not clearly erroneous. We therefore AFFIRM the district court’s judgment for AISD.