Alison Palmer, on behalf of a class of women Foreign Service Officers (“FSOs”), appeals two findings of the district court that Palmer failed to prove allegations of sex discrimination in the United States Foreign Service between *3781976-1983.1 We affirm the district court's finding that plaintiffs did not submit sufficient evidence of discrimination in evaluations of women officers’ potential for years other than 1977. However, we reverse the district court’s finding of no discrimination in promotions of class 5 women FSOs and remand the case on that issue for further proceedings consistent with this opinion.
I. Background
A. The Foreign Service
The Foreign Service (“Service”) is the United States professional diplomatic corps. Upon acceptance into the Service, an officer is assigned to one of four areas of functional specialty called “cones.” The four cones are political, economic, administrative and consular.
The Service operates on an “up or out” formula that obliges those not promoted after serving a specified period at a specified level to leave the Service. Foreign Service Officers serve at ranks from class 8 (the lowest) to class 1 (the highest). See 22 U.S.C. § 3963. Officers begin to compete for promotions from class 6 to class 5; promotions from class 8 to class 7 and from class 7 to class 6 are noncompetitive.
Promotion decisions are made by Selection Boards and are based exclusively on a review of the contents of an FSO’s personnel file. Those files are made up of each individual’s performance evaluations (which include evaluations of potential) while in the Foreign Service, records of training completed while in the Foreign Service, records of awards, letters of commendation, recommendations for disciplinary actions and Inspector Efficiency Reports relating to the officer.
B. The History of this Case
The classes initially certified in these consolidated civil actions consisted of all women FSOs employed by the Department of State at any time between February 4, 1976 and May 5, 1985 and unsuccessful female FSO applicants who applied between February 4, 1976 and June 21, 1983. The complaints filed in 1976 and 1977 alleged sex discrimination against the class in hiring and wide-ranging discrimination in employment practices. The hiring claims were settled by consent decrees entered in 1983 and 1985. The remaining issues proceeded to trial.
At the conclusion of the trial, the district court entered judgment for defendant. Palmer v. Shultz, 616 F.Supp. 1540 (D.D.C. 1985). On appeal, we reversed the district court’s findings of nondiscrimination in assignments, evaluations of female FSOs for potential, awards, and promotions from class 5 to class 4, and remanded for further findings based on the existing record. Palmer v. Shultz, 815 F.2d 84, 116 (D.C.Cir.1987).
On remand, after re-examining the statistical evidence presented at trial, the district court entered findings of discrimination in the areas of initial cone assignments, evaluations for potential in 1977 and Superior Honor Awards.
The district court also found, however, that plaintiffs had failed to prove their case with respect to discrimination in promotions from class 5 to class 4. Additionally, the district court found that plaintiffs had failed to prove discrimination in evaluations for potential for the other years in issue — 1976, 1978-1983.
Appellants argue that the district court erred in failing to give adequate consideration to all the evidence that the Service discriminated in evaluations throughout the period from 1976-1983. Appellants further contend that the district court made both *379legal and factual errors in concluding that the Service had not discriminated in promotions against women FSOs. We address each contention in turn.
II. Analysis
1. Discrimination in Evaluations for Potential
In arguing that the district court erred in finding that appellants had failed to prove discrimination in evaluations for 1976 and 1978-1983, appellants rely primarily on the court’s failure to consider the statistical evidence of discrimination presented for 1977 as evidence of similar discrimination in 1976 and 1978-1983.2 But appellants conceded at argument that there was no agreement between the parties that the statistical information gleaned from the 1977 evaluations was to be considered representative of the other years in issue. Moreover, appellants initially requested discovery of evaluations from two different years but prior to trial acquiesced in receiving evaluations only for 1977. See Reply Brief at 3, fn. 3 (“In light of the short time available prior to trial and the fact that it would take a substantial period of time to perform the survey ..., plaintiffs accepted the single year’s data.”). Consequently, we find that appellants could not expect to rely on the 1977 evaluations as sufficient evidence of discrimination in evaluations for years other than 1977.
In addition to the statistical evidence, however, appellants rely on the testimony of two witnesses which the district court failed to mention and which appellants claim are probative of discrimination in evaluations for years other than 1977.3
First, appellants offered the testimony of Mary Lee Garrison, who served on the FSO Selection Board in 1981. Garrison testified that she discerned a pattern of high performance ratings but low potential ratings for the same women. Joint Appendix (“J.A.”) at 190-91. Second, desk officer Ellen Shippy testified that she personally was evaluated highly in performance but not in potential. J.A. at 169-70. We find this anecdotal evidence to be of minimal probative value and hold that the district court did not err in failing to discuss it.
In sum, then, the lack of statistical evidence for years other than 1977 and the scanty anecdotal evidence offered by appellants justify the district court’s conclusion that appellants failed to prove their case of discriminatory potential evaluations outside of 1977. See Palmer v. Shultz, 815 F.2d at 101 (The factual findings of the district court can be reversed only if they are clearly erroneous in light of all the evidence in the record.).4
2. Discrimination in Promotions
The district court’s finding of no discrimination in promotions from class 5 to class 4 was, however, based on a clearly erroneous interpretation of the defendant’s evidence.5 In our view, the district court could not have reasonably determined that the defense evidence proffered outweighed plaintiffs’ evidence of discrimination in promotions.
For example, after recognizing that discrimination in the granting of Superior *380Honor Awards in class 5 from 1976-1983 “provides some additional support to the plaintiffs’ promotion discrimination claim since awards are also considered by Selection Boards during the promotion process,” Palmer v. Shultz, 662 F.Supp. at 1574, the court went on to reason, erroneously, that the fact that other awards, such as the Distinguished Honor Award, were not awarded discriminatorily during the same period “persuasively supports a finding that promotion discrimination did not occur between 1976 and 1983.” Id. We find this reasoning flawed.
A woman FSO who has been discriminatorily denied a Superior Honor Award is clearly disadvantaged when promotion decisions are made and she finds herself competing against male counterparts who may have received such an award. The fact that she has been judged fairly for purposes of receiving a different award in no way offsets the discrimination she incurred in failing to obtain a Superior Honor Award; she still has one less kudos in her file for the Selection Board to review when promotion time comes along. The lack of discrimination in the Distinguished Honor Award in no way diminishes the nexus between discrimination in Superior Honor Awards and discrimination in promotions.
Similarly, the district judge found that class 5 women FSOs were discriminated against in stretch and downstretch assignments from 1976-1981, which resulted in women receiving more than their proportionate share of downstretch assignments during that period.6 He further found that “it is likely that the opportunity to successfully perform at a position rated higher than an FSO’s grade will help assure a promotion from class 5 to 4 while performance at a lower rated position could indicate a lack of initiative and other negative qualities.” Palmer v. Shultz, 662 F.Supp. at 1574. Nonetheless, the judge determined that since there was no discrimination found in stretch or downstretch assignments for 1982 and 1983, and since stretch/downstretch discrimination was found in classes 4 and 7 (where no promotion discrimination was found), the stretch/downstretch discrimination for class 5 FSOs was not particularly probative of promotion discrimination. We disagree.
That female FSOs were not discriminated against in stretch/downstretch assignments in either 1982 or 1983 in no way offsets the discrimination incurred by female FSOs from 1976-1981. Their records in those years were still discriminatorily tainted, regardless of what their 1981-1983 records displayed. And the fact that class 4 female FSOs were discriminatorily over-assigned to downstretch positions but were not discriminated against in promotions only indicates that stretch/downstretch discrimination does not always overwhelm the other factors on which promotion decisions are based.7
Finally, the district judge found discrimination in the evaluations for 1977 and recognized that “discrimination in evaluations in general is probative of promotion discrimination since evaluations are heavily relied upon by Selection Boards” Palmer v. Shultz, 662 F.Supp. at 1574. The district court nonetheless concluded that “any residuary impact caused by the 1977 evaluations [for years other than 1977] is reduced as subsequent non-discriminatory evaluations are made annually and added to the FSO’s personnel file for consideration by future Selection Boards.” Id. (emphasis added). This reasoning too is flawed.
The precepts used by the Selection Boards to guide them through promotion decisions require the Boards to
place the greatest emphasis on the most recent 5 years of service [but the Boards] should not give undue weight to any single evaluation report.
*381J.A. at 75a (emphasis added). Thus it is clear that the “residuary impact” on promotion decisions of the 1977 evaluations is the same for 1981 as it is for 1977 (or any intervening year); it is not reduced.
In sum, then, a substantial portion of the defense evidence that the district court found rebutted appellants’ evidence of promotion discrimination could not reasonably have been so interpreted. Consequently, we hold that the district court’s finding that defendant’s evidence of lack of promotion discrimination successfully rebutted plaintiffs’ evidence of promotion discrimination was clearly erroneous. We therefore remand the case for a new appraisal of the evidence regarding promotion discrimination. If after reconsidering that evidence, the district court finds that appellants have indeed made out a case that sex discrimination was a substantial factor in the relevant promotions, the Service must either demonstrate that it would have made the same promotion decisions without taking gender into account, or lose the case. See Price-Waterhouse v. Hopkins, 490 U.S. 228, 109 S.Ct. 1775, 104 L.Ed.2d 268 (1989).8
III. Conclusion
Based on the evidence submitted, the district court’s conclusion that plaintiffs failed to prove discrimination in potential evaluations for years other than 1977, was reasonable. The district court’s failure to address specifically plaintiffs’ anecdotal evidence was not a critical error. Consequently, we affirm the district court’s finding of no discrimination in evaluations for years other than 1977. However, the district court's conclusion that defendant’s evidence of nondiscrimination in promotions successfully rebutted plaintiffs’ evidence of discrimination was based on a clearly erroneous interpretation of the significance of defendant’s evidence. Consequently, we reverse the district court’s decision on the promotion issue and remand that portion of the case for further proceedings consistent with this opinion.
Affirmed in part, reversed and remanded in part.
. Two other issues — discrimination in out-of-cone assignments and inadequate relief — were presented on brief by appellants. At argument, however, counsel informed us that those issues were in settlement discussion, and subsequently a provisional decree has been entered in the district court incorporating the parties' agreement. Palmer v. Baker, No. 76-1439 (D.D.C. May 1, 1990) (Order granting preliminary approval of settlement issues). These issues have been withdrawn from the appeal without prejudice to renewal in the event the decree is not finally allowed.
. The district court found that the "statistics presented by plaintiffs demonstrated a statistically significant disparity between men and women in the 1977 potential ratings of FSOs in classes 5 and 6. Therefore, plaintiffs have established a prima faáe case that women in classes 5 and 6 were discriminated against in evaluations during 1977.” Palmer v. Shultz, 662 F.Supp. 1551, 1573 (D.D.C.1987).
. See Palmer v. Shultz, 815 F.2d at 96 ("Plaintiffs in Title VII pattern and practice cases need not rely on statistical evidence alone. Because the ultimate issue in a disparate treatment case is whether the disparity resulted from unlawful discriminatory animus, plaintiffs may introduce any additional evidence which is probative of this issue.").
. The only other evidence that Palmer offered on the subject of discrimination in evaluations was the subjective nature of the evaluation system during the years in issue. The district court took note of this evidence in its findings of fact. Palmer v. Shultz 662 F.Supp. at 1563.
. Consequently, we do not reach appellants’ contention that the district court applied the law incorrectly to appellants' promotion discrimination claims.
. An FSO holds a personal rank. Additionally, each position that an FSO might hold is classified at a certain rank. An FSO can be assigned to a position at, above or below his or her own rank. A position below an FSO’s rank is a downstretch assignment.
. That class 7 FSOs were not discriminated against in promotions is meaningless since promotions from classes 7 to 6 are pro forma.
. It is important to note, however, that the four Justice plurality in Price-Waterhouse required only that plaintiff show that sex discrimination was a "motivating” factor in the employment decision, not that it was a "substantial” factor. See Price-Waterhouse, 109 S.Ct. at 1787-88 ("Once a plaintiff in a Title VII case shows that gender played a motivating part in an employment decision, the defendant may avoid a finding of liability only by proving that it would have made the same decision even if it had not allowed gender to play such a role.”). In separate concurrences, however. Justices White and O'Connor agreed that in order for the burden to shift to the defendant, the plaintiff in a Title VII case must show that gender was a "substantial" factor in the employment decision in issue. See id. at 1795, 1805.