Richard D. Rowe, a black male, filed this employment discrimination suit under Title VII of the Civil Rights Act of 1964, as amended 42 U.S.C. § 2000e et seq., against his former employer, Numerical Control, Inc. (NCI), a manufacturing concern located in Tennessee. 1 While his complaint was somewhat ambiguous and although other issues were raised, Rowe’s major contention was that NCI’s refusal to rehire him, following a layoff, was racially motivated. Judge C. G. Neese, following a one-day bench trial in March 1980, filed a memorandum opinion in which he found that NCI’s refusal to rehire Rowe was not the result of racial discrimination. Rowe appeals from this judgment. For the reasons stated below, we reverse and remand the case to the district court for further proceedings.
I.
On October 25, 1973, Rowe was hired by NCI as a benchman at its Tullahoma, Tennessee plant. At this time, NCI had 9 black employees out of a total of 236 employees. 2 *90 None of these nine individuals occupied a supervisory position. According to Rowe’s testimony, black employees in his department were discriminatorily supervised by the foreman. On one occasion, the foreman referred to the boxer Muhammad Ali as a “nigger.” When Rowe objected to the foreman’s use of that term, the foreman allegedly became angry and, according to Rowe, gave him “difficulty and problems” thereafter. Specifically, Rowe testified that, as a union steward, he encountered difficulty in negotiating with his foreman, although none of his white counterparts encountered a similar problem. On April 2, 1976, Rowe sustained a back injury at work and was granted a medical leave of absence until July 21, 1976, or until released by his doctor.* * 3 But before his medical leave had expired, Rowe was notified on July 9, 1976, that he had been laid off work. NCI had never before laid off an employee who, at the time of his layoff, was on a medical leave of absence.
After Rowe was released by his doctor for return to work in February 1977, he attempted to regain his job at NCI. When he was not recalled, he spoke with the plant superintendent who allegedly made several derogatory comments concerning the EEOC and NAACP in connection with Rowe’s desire to return to work.
In October 1978, NCI began to hire additional employees, and Rowe submitted his application to be rehired. The normal method in hiring new employees consisted of the company’s personnel manager, production superintendent, production manager and production foreman evaluating the applicant’s experience, training, references and appearance. A different procedure was utilized, however, for former employees of NCI who had lost their seniority rights to recall. The names of the former employees were placed on a list which was then sent to each foreman, and the decision as to rehiring these former employees was left entirely to NCI’s foremen. No guidelines or standards were used to govern a foreman’s decision. 4 Nor was a foreman obliged to state, in written or oral fashion, any reason for his rejection of a former employee. It was simply assumed by the personnel manager *91 that the rejection of a former employee was based on the applicant’s previous work record. 5
Rowe, considered by NCI to have lost his seniority rights, 6 had his name included on the list of former employees, both black and white, who were being considered for the available positions. Although it is undisputed that Rowe was qualified for the available jobs, he was not selected by the foremen among whom the list was circulated and, consequently, was not reemployed. Because the number of former employees selected for rehire was not sufficient to fill all of the job vacancies, NCI subsequently hired eight new employees. 7 At least one of the white persons employed to perform a job which Rowe was qualified to perform was considered by the personnel manager to be less qualified than Rowe.
II.
The district court’s opinion does not reveal under what theory of Title VII liability Rowe’s complaint was analyzed. The district court merely concluded that the reason Rowe was not rehired was that he had not been selected by NCI’s foreman from the list of former employees circulated among the foremen. Referring to NCI’s procedure as “a rather odd
modus operandi”
and recognizing “the dangers of abuse inherent in a rehiring system such as this,” the judge nevertheless considered this procedure a “test” that was reasonably related to job performance, citing
Griggs v. Duke Power Co.,
On appeal, Rowe contends that the district court erred by failing to apply the disparate treatment analysis set forth in
McDonnell Douglas v. Green,
*92 III.
A.
As recently articulated by this Court in
Chrisner v. Complete Auto Transit, Inc.,
Either theory may be applied to a particular set of facts.
International Brotherhood of Teamsters v. United States,
The record reveals that there was some confusion on the part of the trial court and the parties as to whether. Rowe’s claims were to be analyzed under the disparate treatment test of
McDonnell Douglas,
under the disparate impact test of
Griggs
or under both. As noted earlier, the district court’s memorandum opinion does not allude at all to
McDonnell Douglas
or the disparate treatment theory and only makes a passing reference to
Griggs.
Nor do the pleadings of the parties offer much assistance in unscrambling this problem.
9
However, after carefully examining the record, we are convinced that the district court erred by confusing these related but significantly different doctrines of Title VII liability.
See, e.g, Williams v. Colorado Springs, Colo. School Dist. # 11,
B.
The procedure used by the employer in the present case permitted the plant fore
*93
men to select, among former employees, those persons they desired to have rehired. The employer thereby delegated the employment decision, as to this group of applicants, to the foremen for their subjective evaluations. This Court has previously noted the problems inherent in selection procedures which rely solely upon the subjective evaluations of all white supervisory personnel.
See, e.g., Senter v. General Motors Corp.,
IV.
Under the disparate impact theory, a plaintiff must produce evidence demonstrating that the employment practice in question selects applicants for employment, reemployment or promotion in a racial pattern which is significantly different from the general pool of applicants.
Albemarle Paper Co. v. Moody,
The test is whether there exists an overriding legitimate business purpose such that the practice is necessary to the safe and efficient operation of the business. Thus, the business purpose must be sufficiently compelling to override any racial impact; the challenged practice must effectively carry out the business purpose it is alleged to serve....
*94
Id.
at 879 (quoting
Robinson v. Lorillard Co.,
In the instant case, the threshold question, with respect to the disparate impact theory, is whether the plaintiff produced sufficient evidence to establish a
prima facie
case in order to require the defendant to demonstrate that the employment practice which produced such an impact is job related. Often such a case is established by the use of statistics demonstrating a disparity in selection rates.
See generally, International Brotherhood of Teamsters v. United States,
The basic problem in the present case, insofar as the statistical evidence is concerned, is that it was totally inadequate to show the impact — much less a substantially disproportionate impact — of the selection process in question. Although plaintiff introduced statistics regarding the racial composition of defendant’s work force, those figures
11
were presumably taken from the Equal Employment Opportunity Affirmative Action Program for the defendant effective August 1, 1976 and reflect data as of August 1, 1976, over two years prior to October 1978 when plaintiff submitted his application to be rehired. More importantly, there are no statistics as to the number of applicants for the available positions in October 1978 and no figures showing how many of those applicants were white and how many were black.
12
With respect to the subgroup of which plaintiff was a member (former employees who had lost their seniority and whose names were placed on a list sent to the foremen), there are no figures to show how many persons were on that list and how many were white and how many were black.
13
On this record there was no evidence on which the trial court could have found that the employment practices of the defendant selected applicants in a racial pattern significantly different from the general pool’ of applicants, and plaintiff therefore failed to establish a
prima facie
case of disparate impact.
Hester v. Southern Railway Co.,
*95
. The trial court concluded, without explanation, that discrimination against the plaintiff “was not shown statistically” and that “there was utterly no showing that blacks and whites were treated differently in the administration of this unusual hiring system.” To the extent these conclusions are considered as findings that plaintiff’s proof failed to establish a
prima facie
case of disparate impact, they are supported by the record and are not clearly erroneous. This being true, it was not necessary for the trial court to consider whether “this unusual hiring system” was job related, because the burden on the defendant of showing that a particular requirement is job related arises only after the plaintiff has made a
prima facie
case of disparate impact. The trial court, nevertheless, stated that, “Despite the dangers of abuse inherent in a rehiring system such as this, the Court cannot say that the test utilized is not job-related. Certainly, it is a reasonable measure of job performance when a former foreman reports that former employees have, or have not, done a good job earlier.” If the plaintiff had succeeded in establishing a
prima facie
case of disparate impact, the record in this case would not have supported any finding that defendant’s unique selection process with respect to former employees was job related. Contrary to the trial court’s statement, there was no report required of the foremen as to their reasons for rejecting plaintiff,
14
and while the personnel manager may have assumed that the rejection of a former employee was based on that person’s previous work experience, the record is silent as to why certain employees were in fact rejected and others accepted. The evaluation of these applicants was left to the exercise of the foremen’s subjective criteria and, as noted above, any procedure employing such subjective evaluations will be carefully scrutinized. If the employer in the present case had been confronted with a
prima facie
case of disparate impact as a result of its hiring practice, the unrestricted discretion and unexplained subjective decisions would have failed to rebut that
prima facie
case.
Brown v. Gaston County Dyeing Machine Co.,
V.
There is an additional and more fundamental problem and one which requires a reversal of the trial court’s judgment: the failure of the court to analyze the plaintiff’s claim under the disparate treatment doctrine.
McDonnell Douglas
established the order and allocation of proof in a private, non-class action alleging employment discrimination. The plaintiff in a Title VII trial must carry the initial burden of establishing a
prima facie
case of racial discrimination. The Supreme Court made it clear that a
prima facie
case is established by a showing that (1) plaintiff belongs to a racial minority; (2) plaintiff applied and was qualified for a job for which the employer was seeking applicants; (3) despite his qualifications plaintiff was rejected; and (4) after plaintiff’s rejection, the position remained open and the employer continued to seek applicants from persons of plaintiff’s qualifications.
In the present case, there is abundant evidence to establish a prima facie case of *96 disparate treatment. Here, Rowe had been previously employed by NCI in the very position that he had applied for. It is also evident that Rowe was rejected and, thereafter, that the available positions for which Rowe applied were filled. At least one of the white persons so employed was considered by the personnel manager to be less qualified than the plaintiff. We hold that, under these circumstances, a prima facie case of disparate treatment was established. 15
Once a plaintiff has proved a
prima facie
case, the burden then shifts to the employer to articulate some legitimate, nondiscriminatory reason for plaintiff’s rejection.
McDonnell Douglas,
While we note that nothing in Title VII abolishes an employer’s managerial prerogatives or its right to refuse to hire undesirable employees, we cannot accept this purported reason. NCI’s reason does not tell the court or the plaintiff whether the employment decision was in fact based upon Rowe’s prior work record or whether the decision in fact was based upon impermissible racial considerations. As noted earlier, NCI’s subjective evaluation procedure effectively delegated the decision to the production foremen. As the district court noted, “there were no standards or guidelines these foremen were required to utilize; each reporting foreman was and is free to decide ‘... for whatever reason ...’ that he does or does not want ... a person listed rehired.” As a result, the record does not disclose the specific reason or reasons why NCI’s foremen did not want Rowe rehired. 16
In
Texas Dept. of Community Affairs v. Burdine, supra,
the Supreme Court clarified the shifting burden analysis of
McDonnell Douglas
and specifically dealt with what is required of an employer in a Title VII case to rebut the presumption of discrimination raised by a
prima facie
case. In order to sustain his burden of production and rebut the presumption, “... the defendant must clearly set forth, through the introduction of admissible evidence, the reasons for the plaintiff’s rejection.”
This obligation arises both from the necessity of rebutting the inference of discrimination arising from the prima facie case and from the requirement that the plaintiff be afforded “a full and fair opportunity” to demonstrate pretext.
In the judgment of this Court, the reason articulated by the employer NCI does not meet the required Burdine standard because it was not substantiated by clear, reasonably specific and legally sufficient evidence. NCI’s stated reason, simply put, explains only that persons to whom the employment decision was delegated did not want him. Rowe’s former foreman may well have made his decision not to have Rowe rehired because of his belief that the plaintiff’s work record was unsatisfactory. On the other hand, he may have made that decision because he personally disliked black people or for any number of reasons. The speculation that is required to determine the precise question is illustrated by the district court’s memorandum opinion which refers to the evidence that Rowe’s foreman had once or twice admonished him to stay at his post of duty and notes that “This could have provided a viable reason for Mr. *97 Rowe’s failure to be requested for rehire the second time around.” While we agree that such a reason could constitute a legitimate reason, no evidence was introduced at trial to demonstrate that it was the reason given by NCI for its refusal to rehire Rowe. As evidenced by the record, the subjective evaluation procedure utilized by NCI did not require the foremen to offer a reason for their rejection of an applicant and, consequently, neither the employer’s personnel manager nor the court itself knew the basis for the foremen’s action. The reference to the occasional admonishments of Rowe do not approach the level of specificity required by Burdine. 17
Equally as important, the failure of NCI to produce such evidence deprived the plaintiff of the required full and fair opportunity to demonstrate pretext. Without an articulated reason presented by the employer, the plaintiff does not know whether the employment decision was made upon his work record or upon an illegitimate racial preference. His offer of proof is somewhat thwarted by this confusion. A plaintiff cannot disprove as a cause for his failure to be rehired a source of dissatisfaction of which he is unaware.
Although plaintiff’s evidence, under the criteria of
McDonnell Douglas,
was clearly sufficient to establish a
prima facie
case of disparate treatment, without regard to the evidence of racial slurs, the trial court erroneously found that the only evidence from which it might be inferred that defendant’s refusal to reemploy plaintiff was racially motivated consisted of the racial slur of the foreman and the derogatory references of the plant superintendent to the EEOC and the NAACP. In considering the evidence of racial slurs in the present case, the trial court considered it in the light of decisions dealing with circumstances involving racially motivated harassment and insult which themselves may sometimes constitute a violation of Title VII. While it is true that unauthorized racial slurs by a foreman could be so excessive and opprobrious as to rise to the level of a violation of Title VII,
Cariddi v. Kansas City Chiefs Football Club, Inc.,
VI.
For the foregoing reasons, the judgment of the district court is Reversed and the case is Remanded for the purpose of entering judgment for the plaintiff upon a determination of the appropriate relief to which the plaintiff is entitled.
Notes
. The action was initially filed against both NCI and United Steel Workers of America, AFL-CIO, a labor union representing NCI’s employees pursuant to a collective bargaining agreement. By a stipulation signed by the parties prior to trial, all issues between Rowe and the union were settled, and the action was dismissed as to the union.
. Prior to trial, the parties stipulated to a number of items including the racial composition of NCI’s work force. These statistics revealed the following pattern in 1976.
*90 Total Occupation Number of Type Employees % Minority Minority Available Employees for job
Technicians 9 3.4
Clerical/Office 10 2.2
Purchasing 1
Craftsmen 2 3.2
Operative 140 6.8
Laborers 16 12.5
Service 2 16.6
. NCI’s personnel manager contended at the trial that the extension “until released by doctor” was a mistake, and that such an extension should not have been given to Rowe.
. NCI’s manager testified at the trial:
A. May I say this, all employees that lost their seniority due to lay-off over the time period, everybody that was on the seniority list, a list was made up, passed to each foreman, they could request that they be rehired or not request. If they were selected then that foreman would ask that they be rehired.
Q. Was Rowe’s name on such a list?
A. Yes, it was.
Q. Was it passed to his foreman?
A. Yes, it was.
Q. Was it reviewed by you and Mr. Garrett (Production Manager) and Mr. Hasty (Production Superintendent)?
A. The foreman that the individuals worked directly for, and all production foremen were given this list, they made the selection in this particular situation on the employees that had lost their seniority.
Q. In other words, you and Mr. Garrett and Mr. Hasty for that purpose, people in the status of Mr. Rowe, didn’t make the decision then?
A. No, we did not.
Q. And Mrs. Millsaps (Coordinator of Affirmative Action Program) had nothing to do with that either?
A. No.
Q. It was strictly done by the foreman on the job?
A. Based on their previous experience with the individual.
Q. Or based on whatever they wanted to base it on, it was their decision to make?
A. Right.
Q. And if I understand it, there was no particular criteria or preference given to Mr. Rowe or others in his place, because they had prior experience, other than submit their names to the foreman?
A. Yes.
*91 Q. And the foreman then can decide, for whatever his reason is, that he does or does not want Mr. Rowe or any others?
A. That’s right.
. The record contains the following exchange between the trial judge and NCI’s personnel manager:
THE COURT: In other words, you wouldn’t really know then for what reason they had for saying, “I don’t want this particular person rehired?”
THE WITNESS: No, we had to believe it was on their work record.
. Rowe contended during the trial that, contractually, he did not lose his seniority right to recall while on medical leave from his employment or for a period of two years after he came off medical leave in February, 1977. He thus claimed seniority as of October, 1978, when NCI began hiring additional employees. NCI disputed this seniority claim and considered Rowe’s seniority as terminating on July 9, 1978, two years after he was given a layoff notice. NCI, therefore, in October, 1978, considered Rowe in the same category as any new applicant for employment, insofar as seniority rights were concerned. The district court considered this claim to be “collateral” to the main contention of racial discrimination and did not resolve it. Rowe has not raised this issue on his appeal to this court.
. At the outset of the trial, the district court read the following statement which was agreed upon by the parties:
The following listed people with seniority dates of hire listed were hired by NCI to perform jobs which the plaintiff, Richard D. Rowe, was qualified to hold and perform. Of these new hirees only one was a minority member.
Linda Barnes January 8,1979
Albert Durkee, Jr. January 22,1979
Allan Kraus January 29,1978
Joe E. Pendleton May 31, 1979
Ralph Russell June 14,1979
James Crittenden June 18,1979
Brian Huffer June 18,1979
Rick German June 19, 1979
. The burden of production must be distinguished from the burden of persuasion which always remains with the plaintiff.
Texas Dept, of Community Affairs v. Burdine,
. As will be discussed infra, there is little doubt that Rowe pleaded a disparate treatment case. However, we believe it arguable that the plaintiff also attempted to plead a disparate impact case. We take note of the following facts:
1. The plaintiff in his complaint alleges that “The defendant NCI has and is presently pursuing a policy, practice and custom of discriminating against and limiting employment opportunities of plaintiff and other black employees solely because of their race.” This implies that the subjective evaluation procedures at issue are an integral component of this practice.
2. The parties entered into a stipulation setting forth statistics regarding the composition of NCI’s work force. See n. 2 supra.
3. Testimony was heard at trial which discussed the number of minority persons hired as a result of the re-selection procedure.
Given these elements and the characterization of the NCI procedure as a “test,” we conclude that the district court felt, as we do, that a disparate impact case was being fashioned.
. As mentioned earlier, there was some confusion as to which liability theory was applicable to Rowe’s claims. The confusion present often exists in Title VII suits. In McDonnell Douglas, supra, the Supreme Court, in outlining the allocation of proofs in disparate treatment cases, interjected a comment concerning the possible application of Griggs to the facts of that case:
We note that the issue of what may properly be used to test qualifications for employment is not present in this case. Where employers have instituted employment tests and qualifications with an exclusionary effect on minority applicants, such requirements must be “shown to bear a demonstrable relationship to successful performance of the jobs” for which they were used, Griggs v. Duke Power Co.,401 U.S. 424 , 431,91 S.Ct. 849 , 853,28 L.Ed.2d 158 (1971); Castro v. Beecher,459 F.2d 725 (CA1 1972); Chance v. Board of Examiners,458 F.2d 1167 (CA2 1972).
. See n. 2 supra.
. The record, sketchy as it is, does indicate that approximately 78 persons were hired and of this number 10 were black. If so, of those hired approximately 13% were black, a percentage in excess of the 1976 percentages available for the jobs in every category except “service.” See n. 2 supra.
. The record does establish that there were white former employees as well as black former employees on the list. The record indicates that six black employees were laid off in 1976, but that three retained their seniority and were recalled by the defendant. The remaining three, including the plaintiff, had lost their seniority and therefore their names were on the list. Of these three, one was selected for reemployment, but he did not elect to return to work. While the record thus shows, at best, that two of the black former employees were not selected as a result of the process, there is no figure as to the number of white former employees who were on the list and also were not selected as a result of the process.
. See n. 4 and n. 5 supra.
. NCI’s counsel, during the course of oral argument, conceded that the record before us sufficiently established a prima facie case under the criteria of McDonnell Douglas.
. See n. 4 supra.
. In
Loeb v. Textron, Inc.,
Although the employer has a burden of production rather than of persuasion, “The employer’s defense must ... be designed to meet the prima facie cases . ..,” International Brotherhood of Teamsters v. United States,431 U.S. 324 , 360 n. 46,97 S.Ct. 1843 , 1867 n. 46,52 L.Ed.2d 396 , . .. and must be sufficient on its face, to “rebut” or “dispel” the inference of discrimination that arises from proof of the prima facie case. [Citation omitted.] A passing reference by just one of many witnesses to some defíciency in the plaintiff’s job rating, for example, would be insufficient. Nor would it be enough to offer vague, general averments of good faith — a plaintiff cannot be expected to disprove a defendant’s reasons unless they have been articulated with some specificity. Cf. Alexander v. Louisiana,405 U.S. 625 , 632,92 S.Ct. 1221 , 1226,31 L.Ed.2d 536 (1972) (averments of good faith in individual selections cannot dispel prima facie case of systematic discrimination).