(dissenting):
Appellant, a GS-9 chemist at the National Institute of Health, seeks advancement to a GS-12 position, and a chance to prove herself as an independent investigator, on the ground that she has been discriminated against in promotion to highly skilled positions because of her sex. I would affirm the trial court’s judgment for Appellee for four reasons. First, even assuming Appellant established a prima facie case of individual sex discrimination by the use of statistics, once the burden shifted to Appellee, it clearly proved that Appellant was not promoted because she simply was not qualified. Second, the statistics introduced by Appellant are not sufficiently detailed to establish a prima facie case of individual sex discrimination in promotions to these highly skilled positions. Third, even assuming the statistics are appropriate in form to establish a prima facie case, they do not come close in substance to establishing a prima facie case of individual sex discrimination. Finally, apart from the statistics, Appellant failed to sustain her burden of proving the qualification element of a prima facie case.
I.
Appellant graduated in 1968 from Duke University with a Bachelor of Science degree. In her science courses, which comprised seventy percent of her classes, she received C’s, D’s, and F’s. Nevertheless, she was hired that year by the National Institute of Health as a chemist with a GS-5 rating. Men and women with better qualifications were hired as GS-7 chemists, and the majority of chemists were women.
Though she thereafter acquired a Masters and Doctorate degree, her promotions came slowly. Dr. Donald Frederickson, one of her superiors, expressed the opinion that she was less sufficient as a chemist than some of her peers. (Tr. 882) In a meeting held on March 18, 1974 to obtain the independent assessment of all the professional staff concerning the chemists in the laboratory, the members of the staff who were familiar with Appellant’s work found that she was average as compared to other chemists. (Tr. 880-81; DX-4)
The record also reflects that when she was working on a project for a Dr. Assman, one of her Ph.D. thesis supervisors, she was unable to obtain usable data while others in the Heart Institute were assaying the same enzyme and doing so successfully. (Tr. 974) Dr. Assman did not recommend her for a promotion, although he recommended two other chemists, including a woman. (Tr. 981-82) Appellant’s superior at the time of the trial, Dr. Pisano, testified that he had told her he rated her the lowest of the three post-doctorate employees that were assigned to him. He further stated she did not spend as much time in the laboratory as other post-doctorate employees. (Tr. 1007-1008)
The evidence at trial also indicates that Appellant was a difficult employee to get along with. This is indicated by the fact that when she started working on her Ph.D. the NHLBI staff thought it necessary to confirm in writing the agency policy that a chemist who achieves an advanced degree is not assured a senior research staff position (independent investigator) when a Ph.D. is *967completed. The memorandum of understanding stated that this policy applied to her. (Tr. 133-34, 146-47, 179, 866-67, PX 95) Previously this policy had been expressed in the form of verbal understandings, but the staff considered that a verbal understanding with Appellant would be impossible. (Tr. 935) In addition, Dr. Frederickson had informed Appellant one month after she had commenced her Ph.D. program that there was no guarantee she would receive the independent investigator position upon completion of the degree. (Tr. 458; DX-11)
Another important fact brought out at trial was that Appellant was not part of the staff fellowship program. In that program, employees with Ph.D.s have the opportunity to increase their skills while proving their developing capacities for innovative creativity.1 The agency rarely offers the independent investigator positions to an inexperienced Ph.D. who is not participating in the program. Even then, not all staff fellows are guaranteed the positions. Edward Nicholas, Director of Personnel at NIH, testified that 75% of NIH senior level positions were filled from within the agency, and these promotions are almost exclusively from members of the program. It is apparent from Appellant’s findings of fact that she refused to enter the fellowship program. By refusing, she is suffering one of the consequences.
These facts clearly indicate that Appellant had frequent personality conflicts with various of her superiors. As concluded by the trial court, her failure to be promoted resulted from this fact, and the fact that she was not as qualified as her peers to take on the heavy responsibility of a high level independent investigator. It is purely incredible to assume, as Appellant would have us do, that the only reason she is not a GS-12 is because she is a woman. Appellee sustained its burden of proving that the promotion decision was based on a legitimate consideration. Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978). Further, Appellant presented no evidence that Appellee’s justification was a pretext for discrimination. Id. at 578, 98 S.Ct. 2943. Thus, the trial court’s conclusion, based on all of the evidence, that Appellee did not violate Appellant’s civil rights under Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C. § 2000e-16, is not clearly erroneous and should be affirmed.
II.
The second reason I would affirm the trial court’s judgment for Appellee is that Appellant’s statistics were inadequate in form to establish a prima facie case of individual sex discrimination to the highly skilled chemist position. This is the principle issue discussed in the majority opinion. The opinion approves of the use of statistics which merely reflect the “minimum objective qualifications” for the positions under review. As support for this test, it cites Internationa] Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977) and Hazelwood School District v. United States, 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977). However, the Supreme Court nowhere establishes this as the standard to be used, and the two cases are clearly distinguishable from Appellant’s. It appears that the minimum qualification needed in the Teamsters case was whether the individuals being considered for hire could drive a truck — a fairly fungible skill. Similarly, Hazelwood involves hiring discrimination of public school teachers where far less skill, training and experience are needed than for chemists doing advanced medical research. Since both cases concerned more common skills and occupations .than research chemists, a litigant’s statistics utilizing minimum objective qualifications for hiring would be much more reliable to establish a prima facie case of an employer’s discrimination.
*968We are here considering research employees engaged in some of the most advanced medical research on the outer horizons of present knowledge. The qualifications for such work should be preeminent. If employees of average ability are filling these positions the public is not getting the quality research that it expects and needs. I would hold that the statistical data introduced by Appellant does not prove a prima facie case. Because of the great degree of skill needed for the job, I would require her initial proof to descend more to particulars and to touch upon the obvious qualifications considered for the job. Just because the qualities demanded are not written down by the employer does not mean they are not considered by it, or are not obvious to the litigants and the Court. As the opinion in the Teamster case states, statistics may be considered as they relate to surrounding facts and circumstances. 431 U.S. at 340, 97 S.Ct. 1843. Similarly, the Supreme Court in Hazelwood School District noted that
[w]hen special qualifications are required to fill particular jobs, comparisons to the general population (rather than to the smaller group of individuals who possess the necessary qualifications) may have little probative value.
433 U.S. at 308 n. 13, 97 S.Ct. at 2742 n. 13. For the Court to accept as a prima , facie case from disappointed chemists statistics that reflect only grade levels, sex, and educational degrees obtained prior to hire is a waste of the Court’s time, and a serious disadvantage to the employers to whom the burden of proof is shifted.
The trial court concluded that Appellant’s statistics were “irrelevant because they include no information regarding qualifications of those males and females available for hiring or promotions.” (App. Vol. I at 23) I would not go as far as the trial court; the statistics are relevant, but are alone insufficient in form to establish a prima facie case of discrimination here because they do not reflect more of the obvious qualifications considered for the jobs. As Appellant’s own expert testified, the statistics do not reflect the pertinent employees’ ages, quality of education, previous work experience, and degrees obtained subsequent to hire at NHLBI.2 (Tr. 705, 736) I also note the statistics do not consider the experience acquired since the various graduate and post-graduate degrees were completed. All of these facts undoubtedly are considered by the employer when determining whether to promote a chemist. It must be stressed that this case demands more particularized statistics for the purpose of proving a prima facie case of sex discrimination because of the uniqueness of the positions being considered. The skills under examination are certainly not fungible, and thus the minimal objective statistics presented by Appellant are insufficient.3
*969III.
Third, even if I were to agree with the majority that the statistics need only reflect the “minimum objective qualifications” for the chemist positions, I fail to see how Appellant’s statistics in substance establish a prima facie case.
Courts have held that in class action cases, discrimination may be proved by statistics. Of course it may, but the statistics must be adequate to make out the prima facie case. In cases of individual discrimination, the use of statistics is less clear. In those cases where the complaining party was a member of a group that was almost totally excluded, Kaplan v. Intern. Alliance of Theatrical and Stage Employees, 525 F.2d 1354 (9th Cir. 1975); Muller v. United States Steel Corporation, 509 F.2d 923 (10th Cir. 1975), the discrimination against the individual is fairly obvious.4 But this is not such a case. The statistical data here is equivocal and inconclusive. Appellant introduced a great deal of data, but much of it points in different directions. From the statistics furnished it may be concluded that women as a group are not discriminated against generally because women constitute most of the employees at the agencies polled. For instance, at the NHLBI, 60.7% of the GS employees in fiscal 1972 were women. In 1974, the figure was 59.4% and in 1976 it was 59.5%. (App. Vol. I at 30-31) In 1974, 55.8% of the professional NIH GS employees who were promoted were women. Therefore, these statistics show that women, as a group, were certainly not excluded from hiring and promotion.
But Appellant’s principal complaint is that women were disproportionately represented in the higher grades. Because the statistics are so inconclusive, it is important to look at them in detail to discern reasons for this disparity, noting that this task should normally be left to the judgment of the trial court as the trier of fact. Hazel-wood School District v. United States, 433 U.S. at 312, 97 S.Ct. 2736. However, be*970cause of the majority’s action in this case, it is necessary to review the evidence.
One of the clearest explanations for the disparity is that there are simply more men than women with the higher degrees to choose from for hiring and promotions. This fact is evidenced by the findings that women only hold 10% of the M.D.s in the United States (Tr. 1049, 1050), and they received only between 14% and 16.3% of the Ph.D.s conferred in select physical and life sciences between 1973 and 1976. (Tr. 763, 764,1047; DX-24) Thus since men numerically dominate the relevant work force it follows naturally that there are more male than female Ph.D.s and M.D.s employed by Appellee.5 Another relevant fact is the finding in Appellant’s statistics that nearly two-thirds of the employees who separated from NHLBI in 1974 were female. Women as a group, therefore, were acquiring less seniority for purpose of advancement. Finally, it is conceivable that women have been hired in greater numbers after the enactment of Title VII in 1964 and the government’s expanded concern over affirmative action. If this premise is true, then the disparity between the number of men and women in the upper grade levels will logically continue for some time until the women as a group gain more experience and seniority.
The most helpful chart introduced by Appellant compares the average GS grade, average length of service, and average age of NHLBI chemists for October, 1975. (App. Vol. I at 137) Unfortunately, she introduced only one such chart. However, the figures it exhibits certainly do not show a clear case of sex discrimination. Of the Ph.D.s, eighteen were men and seven women. The men had an average grade of 14.28 while the women’s average grade was 13.29. This in itself is not dispositive, considering the factors discussed above. In addition, the men had approximately one more year’s service than the women.6 It would be exceptional if the figures for both sexes were identical. For chemists holding only a M.S. or B.S. degree, women fared much better than men. In 1975, although there were almost an equal number of men and women employed with those degrees, the women held a higher average Civil Service grade than the men. The women did have a higher average age and length of service in those groups. However, it would be no more credible for a man to claim sex discrimination on the basis of those figures because, as is stated above, it would be exceptional if they were identical. The only conclusion the trial court should have drawn is that the figures are so close, and the reasons for any disparities so obvious, that Appellant failed to establish a prima facie case.7
*971IV.
Even apart from the statistics, Appellant failed to prove a prima facie case because she failed to prove she was qualified for the job. This proposed holding is different from that discussed in Part I, where it was assumed that even if she had established a prima facie case, the Appellee had in turn sustained its jiurden of proving on rebuttal that she was lot promoted for reasons other than her sex.¡ Here it must be emphasized that Appellant initially carries the burden of proving sje is qualified for the job to first make oit her prima facie case.
In 1973, tje Supreme Court established the necessarj ingredients for proving a prima facie cas in a private, non-class action discrimiratkfi suit in McDonnell Douglas Corp. v. Gren, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.id 68 (1973).
The coiplainant in a Title VII trial must tari} the initial burden under the statute ofistablishing a prima facie case of racial jiscrimination. This may be done by sowing (i) that he belongs to a racial miniity; (ii) that he applied and was qualiid for a job for which the employer Us seeking applicants; (iii) that,' despi' his qualifications, he was rejected; ál (iv) that, after his rejection, the po.ion remained open and the employer ctinued to seek applicants from personof complainant’s qualifications.
Id. at 802, 93 St. at 1824 (Emphasis added). After nog that statistics may be helpful to plaints in proving a pretext for discrimination, i. after the employee has proven a primia.de case and after the employer has induced evidence that the employment decin was based on reasons other than race (¡ex, the Court cautioned that such generaeterminations “may not be in and of therrves controlling as to an individualized hiring decision, particularly in the presence of an otherwise justifiable reason for refusing to rehire.” Id. at 805 n. 19, 93 S.Ct. at 1826 n. 19.
Subsequent Supreme Court cases approving the use of statistics in Title VII cases have usually addressed either class action suits or “pattern and practice” suits brought by the government, neither of which specifically address individual employees. See Hazelwood School District v. United States, 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977); Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977); International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). The Supreme Court’s ruling in Furnco Const. Corp. v. Waters, 438 U.S. 567, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978) clarifies this issue.8 Furnco.is the only post-McDonnell Douglas case which addresses individual discrimination. The Court specifically applies the five-part test' of McDonnell Douglas, and distinguishes Furnco from cases dealing with employment tests, particularized requirements, and “pattern and practice” cases. Id. at 575 n. 7, 98 S.Ct. 2943. The Court then noted that the Furnco employer had conceded for all its purposes that respondents were qualified, thus removing the issue from Supreme Court review. Id. at 576 n. 8, 98 S.Ct. 2943. Therefore, even though Furnco makes it clear that statistics are a relevant form of proof, the five-part test is still applicable.
The logical conclusion is that for suits like Appellant’s, i. e. private, non-class action discrimination suits, the plaintiff must couple the optional use of statistical evidence with proof of personal qualification to make out a prima facie case. The majority opinion fails to specifically address this issue, *972instead, relying on the statistics alone, and the “minimum objective qualifications” to hold Appellant established a prima facie case. For all of the factual reasons expressed in Part I, it is submitted that Appellant has not sustained her burden of proving she was qualified to be an independent investigator.
V.
I would find on the basis of the entire record that the trial court’s conclusions that the Appellee had not engaged in sex discrimination in refusing to promote Appellant had not been shown to be clearly erroneous. In addition to a failure to make out a prima facie case on the basis of statistics, the judgments expressed by her superiors as to her ability, when coupled with the fact that the agency employs women in the agency and as chemists in excess of their ratio in the population strongly supports a conclusion on the merits that sex discrimination was not the reason she failed to be promoted. Rather it was a reflection of her average ability plus her failure to conform to general practices. Most important in this latter connection was her refusal to enter the staff fellowship program, one of the chief means used to select independent investigators.
To my mind this case presents an issue that is of great national concern — is the cry of discrimination going to be used as a means for the promotion of underqualified employees to positions requiring great ability? Chief Justice Burger expressed this very concern for a unanimous court (Justice Brennan not participating) in Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), a case upholding an employer’s use of job testing which reasonably measures job performance.
Congress has not commanded that the less qualified be preferred over the better qualified simply because of minority origins. Far from disparaging job qualifications as such, Congress has made such qualifications the controlling factor, so that race, religion, nationality, and sex become irrelevant.
Id. at 436, 91 S.Ct. at 856. Similarly, are the courts shifting to these employers an unfair burden of proof because of a litigant’s scanty prima facie case based on inconclusive statistics? Employers should resist all unsupported claims less the efficiency of our work force be diminished.
To the extent indicated abore I respectfully dissent from the majority opinion.
. As noted at trial and in the majority opinion, independent investigators must be able to perceive a specific research problem and develop a hypothesis and protocol for determining its truth or falsity. (Tr. 882)
. If the statistics do not reflect degrees obtained subsequent to hire by Appellee, they do not even reflect Appellant’s degrees.
The Fourth Circuit discussed this issue in Roman v. ESB, Inc., 550 F.2d 1343 (4th Cir. 1976) (en banc), where it reviewed the plaintiffs attempt at proving that the employer’s wage scale favored whites over blacks.
The district court examined these contentions and rejected an attempt by the appellants to establish such a disparity by average wages of whites and blacks on a plantwide scale without regard to differences in level of skill, education and training. We agree. Id. at 1355.
Similarly, in Dobbins v. Local 212, International Brotherhood of Electrical Workers, 292 F.Supp. 413, 445 (S.D.Ohio 1968), the Court stated:
It is one thing to presume or assume, prima facie-wise or otherwise, that a significant number of a group have the qualifications for schooling or voting, or jury service. It is another thing to assume, prima facie-wise or otherwise, that because a certain number of people exist, be they Wjhite] or N[egro], that any significant number of them are lawyers or doctors, or merchants, or chiefs — or to be concrete, are competent plumbers or electricians, or carpenters.
See also Hester v. Southern Ry. Co., 497 F.2d 1374, 1379 n. 6 (5th Cir. 1974) (“A more significant comparison might perhaps be between the percentage of blacks in the population consisting of those able to type 60 wpm or better and the percentage hired into the Data Typist position by Southern.”).
. The possible use of more particularized statistical evidence was explained by four economists in Gwartney, Asher, Haworth, Haworth, Statistics, the Law and Title VII: An Economists View, 54 Notre Dame Law. 633 (1979):
*969Since lawyers and judges seldom have been trained in economics and statistics, the legal process has, heretofore, failed to appreciate fully the applicability of earnings determination estimates derived via regression analysis or of analysis of various techniques applied to class definitions and representation questions. Of course, a multiplicity of “skill factors” will generally be relevant to the total evaluation of an employee. Sometimes it will be impossible to fully qualify a “skill factor.” Other times, it will be necessary to use an indirect indicator to measure the impact of a factor. For example, in instances where cognitive skills are important to job performance, years of schooling might be utilized as an indicator of cognitive skills, since data on the latter are not directly available. Fortunately, many potentially important skill factors are quantifiable. Employee data are usually available for skill factors such as (a) years of work experience in related areas, (b) years of vocational training, (c) seniority (work experience with current employer), (d) quantity of schooling, (e) quality' of schooling, (f) typing words per minute and (g) indicators of employee dependability (e. g., a low absenteeism rate). Both economic theory and common sense suggest that these skill indicators will influence the productive contribution of employees, and in turn, their earnings, occupational distribution and even their commonality and typicality under the • definition of a class. When white and black workers possess differing amounts of these skill factors, racial skill differentials are clearly a potential “nondiscriminatory explanation” of the disparity between the unadjusted average earnings or representation of white and black employees and may be helpful in defining a class.
Id. at 655-56.
. For example, the statistics in the Teamsters case exemplify the great disparity between racial classes employed as truck drivers.
As of March 31, 1971, shortly after the Government filed its complaint alleging systemwide discrimination, the company had 6,472 employees. Of these, 314 (5%) were Negroes and 257 (4%) were Spanish-surnamed Americans. Of the 1,828 line drivers, however, there was only 8 (0.4%) Negroes and 5 (0.3%) Spanish-surnamed persons, and all of the Negroes had been hired after the litigation had commenced. With one exception — a man who worked as a line driver at the Chicago terminal from 1950 to 1959 — the company and its predecessors did not employ a Negro on a regular basis as a line driver until 1969. And, as the Government showed, even in 1971 there were terminals in areas of substantial Negro population where all of the company’s line drivers were white.
International Brotherhood of Teamsters v. United States, 431 U.S. at 337, 97 S.Ct. at 1855. (Emphasis in original) In such a case, statistics are valid.
. In 1974, NHLBI’s professional staff included 52 Ph.D.s and 9 M.D.s who were male, and 14 Ph.D.s and 3 M.D.s who were female. (App. Vol. I at 141) The figures for 1975 were 63 Ph.D.s and 10 M.D.s who were male, and 17 Ph.D.s and 3 M.D.s who were female.
. The length of service is important when tied in with the enactment of Title VII of the Civil Rights Act of 1964. The average length of service for men and women chemists at NHLBI in 1975 was 15.6 and 14.9 years respectively. Since, most of the chemists were employed prior to the enactment of Title VII and most prior to that time were men, it would be logical that there are more men in these positions at the present time.
This situation contrasts with the Teamsters case, where the Supreme Court found the pattern of racial discrimination to continue in the employer’s hiring practices long after Title VII was enacted. 431 U.S. at 341, 97 S.Ct. 1843. The employer had argued the disparities existed because of a pre-Act imbalance and post-Act low personnel turnover. The Court stated this “argument would be a forceful one if this were an employer who, at the time of suit, had done virtually no new hiring since the effective date of Title VII.” Id.
. This case bears much similarity to Roman v. ESB, Inc., 550 F.2d 1343 (4th Cir. 1976) (en banc) where the Court held that the statistical evidence introduced at trial did not prove racial discrimination in hiring and promotions. ESB had an overall non-white employee percentage of 54% prior to a layoff, which was well above the percentage of the black population in the community.
Overall, whites do outnumber blacks in the higher job classifications and blacks outnumber whites in the lower classifications, but any imbalance is not striking. .
The spread here of blacks through the labor grades is not such as would alone create a prima facie case of discrimination. This is *971especially true ight of the highly skilled work done by t\jgher pay grade workers such as tool and.
Id. at 1353-54.
. The Furnco Co also clarified that a McDonnell Douglatma facie case only raises an “inference” oknmjnation because the Court presumes th;ts, if otherwise unexplained, are “more likely than not based on the consideration of impermissible factors.” 438 U.S. at 577, 98 S.Ct. at 2949. This finding by the Court is- not the equivalent of an ultimate finding of fact.
For more detail on the two kinds of prima facie cases, see J. Wigmore, Wigmore on Evidence § 2494, at 293 (3d ed. 1940).