WILLIAM E. DOYLE, Circuit Judge.
Plaintiffs-Appellants filed an action in the District Court for the District of New Mexico in which they sought to recover damages from Appellee, Sandia, a corporation which is wholly owned by the United States, and contracts with Energy Research [1385]*1385and Development Administration. Sandia operates a multi-program laboratory; its responsibility is to carry on engineering research and development of nuclear ordinance.
In 1973 Sandia reduced its work force, at which time the seven plaintiffs were discharged or laid off. The claim of plaintiffs was and is that Sandia discriminated on the basis of age; that it violated the Age Discrimination Act, 29 U.S.C. §§ 621 et seq.
The plaintiffs were employed in the following capacities:
Plaintiff Skolnick was a staff associate technician. He performed design and development of diagnostic procedures.
Plaintiff Kinney was a commercial and technical artist.
Plaintiff Smith was a geologist.
Plaintiff Souther was in computer programming and data processing.
Plaintiff Cianciabella was in project engineering.
Plaintiff Breehl did design and development work in the wire and cable department. His expertise was in the flexible circuit field.
The present appeal is part of a larger age discrimination case in which Sandia was defendant. The original suit was by employees of Sandia. It was consolidated with an action brought by the Secretary of Labor for the benefit of a large number of employees. The cases were tried first on the issue of liability, and second as a remedial phase.
In the liability case the court considered the termination of employees during the company’s 1973 layoff. An Interlocutory Opinion was entered in which the court concluded that the evidence established a prima facie case that a pattern and practice of age discrimination existed in the selection of employees laid off who were between the ages of 52 and 64.1 On that occasion the court concluded that the evidence did not show a prima facie case of age discrimination in the lay-offs of persons between the ages of 40 and 51.
The remedial phase of the case consisted of a series of trials in which a number of individual claims were considered. With the exception of one claim all of the trials were concerned with employees who were between 52 and 64 years of age. They fell within the group of employees which the court found had established a prima facie case of age discrimination. As to the employees between 40 and 51 years of age, Sandia’s motion to dismiss was granted with regard to the group represented by the Secretary of Labor. The finding was that Sandia had prevailed against the Secretary’s pattern and practice allegations as to these employees. The court did not grant the dismissals as to the individuals between 40 and 51, who were not represented by the Secretary of Labor. The court stated that the first phase of the trial had not been completed as to those plaintiffs. Thus, the ruling was that they were entitled to a trial of both the liability and remedial phases of their claims.
In January 1980, following a trial, the court entered individual findings and conclusions as to all seven of the present plaintiffs. The court entered a Partial Final Judgment on June 11, 1980 dismissing the lay-off and salary claims of each of the plaintiffs. The present appeal is by the seven plaintiffs from the partial final judgment of June 11, 1980, together with certain determinations stated in the October 20, 1977 Interlocutory Opinion.
Ruling of the Trial Court.
The private plaintiffs between the ages of 40 and 51 at the time of the lay-offs alleged both a “pattern and practice” of discrimination on the part of Sandia, and also individual discrimination. The court [1386]*1386stated that while the liability trial pertained to the “pattern and practice” issues of the plaintiffs’ cases, it did not include a hearing on their claims of intentional discrimination. The holding was that the seven individual plaintiffs were entitled to a trial on the merits of their ease.
The trial on the merits was held in January 1980. The trial court imposed the burden on each plaintiff to establish his prima facie case of age discrimination by showing:
1. He was within the protected age group;
2. He was doing satisfactory work;
3. He was discharged despite the adequacy of his work;
4. His position was filled by a younger person.
The trial court declared that if a plaintiff established a prima facie case, the burden would shift to the defendant to show that the termination was attributable to factors other than age; plaintiff was then authorized to show that the reasons articulated by Sandia were a pretext to cover up discrimination.
The trial court found that Sandia had succeeded in articulating reasons other than age for the terminations; that the jobs in question had not been assumed by younger persons; that plaintiffs had failed to show statistical evidence in support of their claim of age discrimination; that Sandia had legitimate, non-discriminatory reasons for the plaintiff’s involuntary termination; and that the reasons articulated by Sandia were not pretextual.
I. Proof of the Prima Facie Case.
Plaintiffs’ contend that the court’s reason for reaching its result was based entirely on the factor that none of the jobs were filled by persons younger than the plaintiffs, and thus a prima facie case had not been shown.
The Supreme Court in the case of McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), laid down four factors similar to those used by the trial court here. However, the listed factors are not to be applied in a rigid manner in determining whether a prima facie showing had been made. Rather the standards are to be adapted to the fact situations being considered. McDonnell Douglas, supra, n. 13; Laugeson v. Anaconda Co., 510 F.2d 307 (6th Cir. 1975); McCorstin v. United States Steel, 621 F.2d 749 (5th Cir. 1980). A particularly clear statement is found in Texas Department of Community Affairs v. Burdine, - U.S. -, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981). The Supreme Court in that case applied the McDonnell Douglas test to a claim of sex discrimination in which there was a failure to promote. The Court did not require a member of a non-protected class to have been hired to fill the position sought by the plaintiff. The Burdine decision emphasized that the failure to promote was under circumstances which gave rise to the inference that there had been unlawful discrimination; that the plaintiff’s failure to be promoted was more likely than not the result of impermissible factors.
If in the present case the sole reason for ruling against the plaintiffs was that the jobs had not been shown to have been filled by younger persons, the plaintiff would have a strong basis for claiming error. The court, however, went forward and determined that Sandia had met its burden of rebutting the prima facie case of the plaintiff. The court then considered whether plaintiffs had rebutted Sandia’s showing of legitimate non-discriminatory reasons, by establishing that Sandia’s explanations were a mere pretext. Considering the fullness of the express findings, it cannot be said that the trial court’s ruling was based on a finding that there was a failure to establish a prima facie case.
II. The Question as to Sufficiency of the Non-Discriminatory Reasons
Once the plaintiffs establish their prima facie cases, Sandia has the burden of [1387]*1387producing non-discriminatory reasons for the terminations. Texas Department of Community Affairs v. Burdine, supra. If Sandia succeeds in satisfying this burden the plaintiffs must demonstrate that the reasons articulated by Sandia were pretextual. Id. If Sandia is able to articulate reasons which are not a pretext for age discrimination, it is not to be held liable for age discrimination. Id.; Williams v. Colorado Springs 641 F.2d 835 (10th Cir. 1981).
It is plaintiffs’ position that the court erred in finding that Sandia had a legitimate reason for terminating the plaintiffs’ employment.
The standard of review is whether the trial court’s decision was clearly erroneous. Reversal requires that the reviewing court so conclude. Zenith Radio Corp. v. Hazeltine, Inc., 395 U.S. 100, 123, 89 S.Ct. 1562, 1576, 23 L.Ed.2d 129 (1969); United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 541, 92 L.Ed. 746 (1948); Ray v. Safeway Stores, Inc., 614 F.2d 729 (10th Cir. 1980). It is the record as a whole which will approve or disapprove the contention of insufficiency. Id.; Thornton v. Coffey, 618 F.2d 686 (10th Cir. 1980.)
Plaintiffs’ argument is- that at the 1977 “liability trial,” there was overwhelming evidence of age discrimination against employees in the 40 to 51 age group. Plaintiffs maintain that there was a prima facie case established which called for a judgment as to the present age group along with that which was entered with regard to the employees in the 52 to 64 age group. We have previously noted that the trial court went beyond the question of evidence of a prima facie case, and went on to consider the succeeding two steps in the three-step procedure for proving discrimination, so that question has become academic. Our concern is with the issue of whether Sandia has articulated a legitimate reason for the terminations, and whether the plaintiffs have been successful in proving that the reasons articulated as to each plaintiff were a mere pretext to shroud discrimination.
A. Consideration of the evidence of Sandia with respect to reasons for the terminations.
The following evidence was given by Sandia in explanation of the terminations.
1) Plaintiff Cianciabella, Age 51 at the time of lay-off: Sandia has said that the reason for lay-off of this man was due to his lack of technical competence, the difficulties he had getting along with others, plus his general abrasiveness. Mr. Cianciabella’s department manager, John D. Kennedy, testified that he was selected for layoff because he was approximately the lowest ranked employee in performance ratings.
2) Plaintiff Souther, age 47 at the time of termination: Sandia claimed that the quantity of Souther’s work was not equal to that of his peers, and that he had failed to complete his assigned project. Souther was also described as abrasive in his dealings with customers. He was the lowest ranking member of his group in terms of performance. Sandia’s director of computing, Lee Hollingsworth, testified that he had recommended termination of Souther on the basis of his unsatisfactory performance records.
3) Plaintiff Kinney, Age 51 at time of lay-off: Sandia maintained that Kinney was not flexible, and was difficult to work with. Sandia further claimed that the department Kinney worked in was a support service within the company which management felt could be cut back drastically. To that end, the Company looked specifically to the “lower performers” as indicated by the performance scores when determining who was to be laid off. Specifically, Kinney’s last three performance ratings were the cause for his lay-off according to Arland Pepjueller, Department Manager.
4) Plaintiff Breehl, Age 50 at the time of lay-off: Sandia claimed that Breehl had declined in his performance level compared [1388]*1388with other employees, primarily because he tended to irritate those with whom he worked. According to Breehl’s Department Supervisor, Charles Tapp, it was due to Breehl’s downward move on the performance chart that he was selected for lay-off.
5) Plaintiff Skolnick, Age 5VA at time of lay-off: Sandia alleged that Skolnick’s work as a staff associate, while satisfactory in quality, was not satisfactory in quantity. It was felt that Skolnick was not versatile, could not do component design, and that his work could be handled adequately by other staff members. Skolnick was ranked the lowest of all staff associates in terms of performance.
6) Plaintiff Stang, age 46 at time of layoff: Sandia claimed that Stang was laid off because his experience was in the manufacturing area, rather than in the engineering area, and that work in the manufacturing area was being reduced.
7) Plaintiff Smith, age 47 at time of layoff: Sandia maintained that with the decline of the war in Southeast Asia, Sandia no longer had need for two soil scientists/geologists. The company decided that the other geologist, Mr. Culp, who was older than Smith, was better qualified, since he had a Ph.D (while Smith did not), was more effective in organizing projects, communicated more effectively, and interreacted better with people than did Smith. Sandia noted that Culp had taken a 2lh year leave of absence from Sandia, without pay, to get his Ph.D.
B. Scrutiny of the evidence.
In the earlier Sandia case, Equal Employment Opportunity Commission v. Sandia, No. 79-1589, Aug. 13, 1980, the trial court had determined that Sandia’s performance evaluation system was totally subjective and was also totally inherently biased against employees in the protected age group. The trial court found that a prima facie case of discrimination had been established as to employees in the 52 to 64 age group, but that it had not been established as to the employees in the 40 to 51 age group. This latter finding, however, does not undermine the suspect nature of the performance ranking system. Therefore, Sandia’s determination to lay off employees between the ages of 40 and 51 solely because of the mentioned subjective ratings is to be carefully considered in deciding whether Sandia had a non-discriminatory reason for terminating those employees.
Sandia’s witnesses testified that plaintiffs Cianciabella, Souther, Kinney and Breehl were terminated as a result of their low performance ratings. This testimony must be judged in the light of our determinations in the review of the 1977 trial that:
1) The performance evaluations were not based on identifiable criteria related to quality or quantity of work or on particulars of performance.
2) Sandia overgeneralized in an arbitrary way that older employees did not have ability to keep abreast of new developments; Sandia’s management employees tended to stereotype older employees and to mark them as unproductive and as becoming technically obsolete;
3) The performance ratings are related to the employee’s job assignments and motivation. So when employees were laid off because they had been in one line of work too long and were rated less versatile than younger employees, there was circumstantial evidence to infer that age bias and age biased policies were at work to produce lower performance ratings.
The evidence which was presented in the recent trial does not dispel our concerns as to age bias finding its way into the performance ratings. That evidence tends to illustrate that age bias is reflected in the low performance ratings of the above four employees. Breehl’s division supervisor said that Breehl’s decline in ratings was attributable to the increasing number of high technology people in the department rather than deterioration of performance on the part of Breehl. Cianciabella’s supervi[1389]*1389sor stated that the performance ratings within his division had to be interspersed with other divisions and the department, and the resultant “variance” meant that the performance ratings numbers “don’t have a lot of weight to them.” Souther’s supervisor testified that one of two factors contributing to Souther’s decline in performance rating was that the organization itself was becoming stronger and more competitive, and that when Sandia acquired new people with greater capabilities than those of their original employees, performance rankings could decrease without decrease in performance on the part of the original employees.
In Mistretta v. Sandia, 639 F.2d 588 (10th Cir. 1980) we noted that evidence adduced at the liability stage of the trial was properly considered at the remedial stages of the trial as well. Thus, the evidence must be considered as a whole in determining finally whether the plaintiffs have succeeded in establishing that they have been the victims of discrimination. Sandia’s contention that unless the plaintiffs have submitted new evidence at this trial which proves the existence of discrimination there has been a failure of proof is not correct and is rejected. The evidence at the earlier trial which raised an indication of inherent bias in the performance rating system is not dispositive of that issue here. However, Sandia has reinforced the evidence through the testimony of the supervisors as to deterioration in performance rankings through no fault of the four plaintiffs.
We are aware of the Supreme Court’s decision in Texas Department of Community Affairs v. Burdine, supra, wherein the court rejected the theory that the defendant must prove the existence of objective factors upon which the terminations of the plaintiffs were based. Thus, subjective factors are permissible if they are non-discriminatory and are not pretextual to cover up discrimination. That is, in presenting legitimate, non-discriminatory reasons for lay-offs, those reasons may be subjective, as long as the reasons cited are not mere cover-ups for discrimination. Also, we are aware, as previously stated, that the trial court’s decision is to be upheld unless it is clearly erroneous. In the case of plaintiffs Cianciabella, Souther, Kinney and Breehl there is no choice but to conclude that the trial court’s conclusions were clearly erroneous.
The record is plain that Sandia’s management acknowledged expressly that the four plaintiffs named above were not terminated primarily because of the enumerated, subjective, non-discriminatory reasons, that is lack of ability to get along with others, lack of flexibility, etc., but were terminated because of their performance rating.2 Although performance ratings are basically objective in nature from their very character, those performance ratings were found to be inherently discriminatory in our earlier opinion; the evidence in the case at bar, which illustrates that performance ratings may decline through no fault of the employee, does not tend to rebut the inference of inherent bias within the performance ranking system. On the contrary, the evidence serves to reinforce our earlier premise that the ranking system is, in fact, inherently biased against older employees. There is plenty of background in the record as a whole to support this.
We do not quarrel with the enumerated subjective factors. Had the termination of the four above named plaintiffs been based, in fact, primarily or solely upon those factors, our present conclusion would be different. In light of the fact that Sandia’s management personnel has testified that those plaintiffs were terminated due to their performance ratings, however, we cannot reach a contrary conclusion and hold that those four plaintiffs were terminated solely because of non-discriminatory subjec[1390]*1390tive factors. Nor are we able to find any evidence to support a finding that the performance ranking system does not contain built-in elements of age discrimination.
Therefore, our conclusion with respect to the four plaintiffs above named and here under consideration is that the trial court’s determination that there was no discrimination was in error. The evidence which was presented at the 1977 trial dealing with the performance tests and the evidence during the present trial explaining the mode in which performance rankings may decline, serves to illustrate that the performance ranking system is inherently biased against older employees. In view of this, it cannot be said that employees terminated from their positions because of low performance ratings were terminated solely because of non-discriminatory reasons.
In the case of the fifth employee of the group in question, namely Mr. Skolnick, the evidence does not show that he was laid off solely because of his performance rating. In his case it does appear that his performance rating was a critical factor in his being selected for lay-off, however, since it is apparent from the evidence that the fact that Skolnick was ranked the lowest of all staff associates in terms of performance was a factor which was relied upon when the decision was made to terminate him from his position. Furthermore, Skolnick’s department head had testified that the performance ratings related more to “mechanics” than to the quality of Skolnick’s work. However, the testimony of Skolnick’s department head was that the lay-off decisions in his department were made on the basis of job performance, job functions and length of service. Thus there was some documentation in support of the conclusion that Skolnick’s termination was not based on his performance rating.
In the light of the fact, therefore, that there was no express evidence in the record that Skolnick was chosen for lay-off because of his performance rating as there was with regard to the four plaintiffs discussed above, we must defer to the trial court’s determination that Skolnick was chosen for lay-off on the basis of the subjective, non-discriminatory reasons elucidated by Sandia, namely that Skolnick's work was not versatile and that the quantity of his work was insufficient.
In contrast to the evidence submitted with respect to plaintiffs Cianciabella, Breehl, Kinney and Souther, the enunciated, non-discriminatory reasons given by Sandia for Skolnick’s discharge are not expressly contradicted by Sandia’s own management personnel. Accordingly, there is a difference between Skolnick’s case and the cases of the first four plaintiffs considered and we conclude that the trial court’s decision with regard to Skolnick is to be affirmed. It was not clearly erroneous.
The same result must be reached in the cases of Smith and Stang. In those cases Sandia did articulate legitimate, non-discriminating reasons for the lay-offs of those two employees. It is true that one factor in their lay-off was declining performance ratings; however, this was not proven to be the motivating force behind those lay-offs. Sandia’s evidence showed that both plaintiffs were laid off because of a severe decline in the particular work which they were qualified to perform, and plaintiffs failed to rebut the evidence and show that the reasons enunciated by Sandia were pretextual. Hence, the trial court was correct in its ruling that Stang and Smith were not discriminated against by Sandia because of their age.
III. Damages growing out of the Stretch Out Policy.
Sandia’s “stretch out policy” involved use of a chart which specified time intervals between salary increases of minimum percentages. Employees were classified within age groups: 29 and under, 30-35, 36-45, and 46 and over. As a result of literal use of the chart, older employees were required to wait longer for minimum salary increases. The trial court held in its 1977 Interloc[1391]*1391utory Opinion that a prima facie case of age discrimination resulting from the stretch out policy had been established against employees in the entire protected age group.
A partial final judgment with respect to the stretch out policy was entered on September 17, 1979. This was with Sandia’s consent. This judgment approved and confirmed a settlement entered into by the Secretary of Labor and Sandia for stretch out claims. Pursuant to that settlement five of the seven plaintiffs here accepted compensation from Sandia. These were Breehl, Souther, Skolnick, Smith and Stang. In its partial final judgment of June 11, 1980, the trial court held that the amounts recovered by the aforementioned plaintiffs were adequate compensation for the losses suffered, and that plaintiffs Kinney and Cianciabella had failed to prove that they had suffered any losses because of Sandia’s stretch out policies. Plaintiffs disagree with this.
It is Sandia’s contention that once the Secretary of Labor settled the stretch out damage claim against Sandia, that claim was settled for all the named employees, and the individual plaintiffs named above may not raise the issue again. This was the result intended according to Sandia and it was the effect of the final judgment which was entered on September 17, 1979. The settlement agreement reads in pertinent part as follows:
II. Ordered, Adjudged and Decreed that defendant be and it is hereby enjoined from withholding payment of lost earnings and interest thereon computed at the rate of 8% per annum in the total of $1,701,790 which the court finds is due under the Act to defendant’s employees named in Exhibit “A” attached hereto as a result of defendant’s salary administration procedure known as “stretch out.” The payment of the $1,701,790 by defendant, which includes social security and income tax deductions, shall fully satisfy all monetary damage claims of the plaintiff pertaining to defendant’s salary administration procedure known as “stretch out” including interest and benefits.
The foregoing language is plain in stating that the claims of the five above plaintiffs who were included in the Secretary’s settlement and are fully satisfied. The effort to relate the present claims to those which were considered in Mistretta v. Sandia, supra, fails for the simple reason that the plaintiffs in Mistretta were not covered by the Secretary’s settlement. The court allowed the Mistretta plaintiffs to prove and recover damages individually in September 1978. This occasion was prior to the settlement by the Secretary. Hence, we are in agreement with the determination of the trial court that plaintiffs Breehl, Souther, Skolnick, Smith and Stang have no further claim to relief for stretch out damages. We, of course, agree with the findings of the trial court that the four plaintiffs were compensated for stretch out damages.
The claims of plaintiffs Kinney and Cianciabella, who were not covered by the Secretary’s settlement, must be considered on their individual merits.
It will be recalled that the trial court held that these two did not suffer, i. e. were not injured, by the stretch out policy. We must view the evidence on this.
Kinney’s personnel record confirms the lack of stretch out damages. In May of 1964 Kinney was receiving $8,100. He was then 43. He received an increase to $8,520 in May of 1966. Every two years until 1973 he received a similar increase. At the time of his discharge in 1973 he was receiving $10,700. He was then 51. During this time the policy of raising Kinney’s salary on a regular basis remained constant. The policy followed in his case did not show that his raises were less frequent when he moved from one category on the chart to another. The raises did not become less frequent when he moved from one age group to another. In view of Kinney’s record it cannot be concluded that the trial court erred [1392]*1392in holding that Kinney was not entitled to stretch out damages.
Plaintiff Cianciabella testified that he had been told by his supervisor that “you’re not getting a raise this year, you’re on stretch out. But your work is good.” However, the record does not reflect that. Mr. Cianciabella’s personnel record going back to 1965 indicates that from May, 1965 until August, 1966 his salary was $8,040. In August 1966 it was increased to $8,520. Fifteen months later, when he was 47, he received a raise to $10,100. In May, 1970, approximately 15 months later, when he was 48, he received a raise to $11,100. From that time on until his termination on March 23, 1973 he did not receive a raise. Thus, until the last year, Cianciabella received regular increases at the ages of 43, 45, 47 and 48; the last two raises occurring when he was over 46. Therefore, the evidence fails to establish that this plaintiff was injured by the stretch out policy. We must, thus, agree with the trial court that this plaintiff was not injured by stretch out.
The trial court’s Partial Final Judgment, entered June 11, 1980, is affirmed in part and reversed in part. The cause is remanded to the trial court for further proceedings consistent with this opinion.