Appellant Namenwirth was denied tenure at the University of Wisconsin in Madison. After the Department of Labor found that she had been denied tenure because of her sex, she brought this Title VII *1237action in federal court; by agreement, the matter was heard by a magistrate, who found that the university had not discriminated against Namenwirth. We affirm.
I.
Marion Namenwirth was hired as an assistant professor by the Department of Zoology of the University of Wisconsin-Madison in September of 1971. She was the first woman hired in a tenure-track position in thirty-five years, and apparently she was the first person in a tenure-track position ever to be denied tenure by the Department.
Her initial contract was for three years, and it was then to be renewed yearly until she was considered for tenure. A faculty member in such a tenure-track position who is not promoted to the tenured rank of associate professor after the sixth year must ordinarily leave the University after her seventh year.
The University of Wisconsin has had a record of sex discrimination. In 1970, prior to Namenwirth’s hiring, only 150 of 2000 faculty members at the University of Wisconsin at Madison were female. Of these, half were in traditionally female departments such as nursing and home economics. Nearly all were in the lower ranks. Fifty academic departments had no women faculty members at all, even though between 10% and 38% of the Ph.D.s being awarded in those disciplines were being awarded to women. Responding to a complaint from a campus action group, the Department of Health, Education and Welfare in July of 1970 began an investigation into sex discrimination at the University. HEW found that there was a pattern of discrimination against women, and found significant discrepancies between the salaries of males and females in the same positions. In spite of an affirmative action program adopted by the university under pressure from the federal government, the response of the science departments was slow. The magistrate found that in 1978, only six of 323 tenured faculty members in the science departments were women.
The record of the Zoology Department was apparently like that of the other science departments. Before Namenwirth was hired, the only woman to hold a tenure track position in the Zoology Department was Dr. Nellie Bilstad. Dr. Bilstad had received her Ph.D. from the University, and had been hired in 1931. The Department granted her tenure in 1942, and promoted her to associate professor in 1950. She remained an associate professor until 1969, the year of her retirement, when she was promoted to full professor. The magistrate found that Bilstad had been treated by her colleagues in a patronizing manner. He found that men who were junior to her and who had done little in the way of research rapidly moved beyond her in both salary and rank. He found that Dr. Bilstad was the victim of purposeful discrimination which reflected the discrimination prevalent in the University and in higher education in general in the time prior to the 1972 extension of Title VII to universities.
In 1971, Namenwirth was the only woman in a tenure track position in the Department. In 1976-77 she was considered for tenure.1 The usual procedure is for the Department’s Salary and Promotion Committee to consider the applicant and make a recommendation to the Department. The committee in this case favored Namenwirth’s promotion by a slight majority but did not recommend promotion to tenure, a fairly unusual procedure. Instead, it prepared a report which, among other things, expressed concern about the strength of Namenwirth’s research and publication record. This report ultimately became part of the Department’s tenure packet on Namenwirth to the Biological Division Executive Committee, the committee responsible *1238for making a recommendation one way or the other to the Dean of the College of Arts and Letters.
When the Namenwirth candidacy was presented for a vote, the Department voted 11-10 against recommending her for tenure. Dr. Namenwirth asked for reconsideration. She was asked to speak to the Department on her own behalf; after she spoke, another vote was taken, the result of which was 12-10 in favor of recommending for tenure. For some reason, reconsideration was asked again; the outcome was negative. Dr. Namenwirth requested reconsideration once more, and the final vote was 11-10 in favor of tenure, with three abstentions.
In preparing for the Department’s consideration of Namenwirth, the Salary and Promotion Committee had requested letters evaluating Namenwirth’s work from scholars outside the department. These letters and other materials relating to research, teaching and committee work were made part of a packet to be sent to the Divisional Executive Committee along with the recommendation for tenure. The packet also included a discussion of the departmental debate, and of the concerns the Department had about the wisdom of granting Namenwirth tenure. It mentioned the closeness of the Department vote. It analyzed the letters from the outside scholars, and pointed out the concerns those scholars had with Namenwirth’s research.
Largely because of the closeness of the vote and the equivocal nature of the Department’s recommendation, the Divisional Committee voted to deny her tenure, and the Dean of the College of Letters and Science, accepting the Divisional Committee’s recommendation, notified the Department that he would not recommend Namenwirth for promotion and tenure. The Department then insisted that the Division clarify its reasons for denying tenure; the Division, under some pressure from Namenwirth, said that it would reconsider the case if the Department would reconsider it first, resubmitting the case with updated documentation. The Department refused, insisting again on a better statement of reasons for the denial. The Division refused to issue a new statement and told Namenwirth that there would be no reconsideration of her tenure denial.
Namenwirth appealed to the University Committee in the summer of 1978. The University Committee recommended that the Divisional Committee reconsider the Namenwirth case, using the documents originally submitted by the Zoology Department and new material to be submitted by Namenwirth herself. They also recommended that she be kept on the faculty one more year.
In October, 1978, the Biological Division Committee reconsidered the Namenwirth decision, and again voted against tenure. The Dean of the College, again following the recommendation of the Divisional Committee, declined to recommend her for tenure, and at her request provided her with the following list of reasons:
(1) The narrow majority of the Zoology Executive Committee recommending you for tenure after initially voting against tenure by an equally narrow majority.
I have serious doubts about the wisdom of granting tenure unless a substantial majority of the executive committee recommends it. Unlike the Zoology Department, a number of our departments require a two-thirds majority to recommend tenure, a provision I favor and recommend to all L & S departments.
(2) The negative reviews by the Biological Sciences Divisional Committee, which twice recommended against your tenure by a large margin.
As you know, the four Divisional Committees are the general faculty’s major quality control mechanism for faculty appointments to tenure. Each Committee operates across the entire University in its specialized area. A Dean must consult the appropriate Divisional Committee on the qualifications of all candidates *1239for tenure. Because the Divisional Committee reflects the opinions of scholars outside the Department but from fields closely related to it, and because it is a disinterested body concerned with the overall quality of the faculty of the University, a Dean must consider its advice on tenure cases very seriously.
(3) My concern about the serious budgetary problems of the College and the likelihood of further budgetary difficulties if our enrollment should level or decline in the next few years, as seems quite possible.
We presently face the need to reduce our staffing in some areas of the College. Under the circumstances, I believe it is especially important that we grant tenure only to those faculty members whose qualifications are beyond question and whose service is likely to be needed regardless of any enrollment fluctuations we can foresee at this time.
A number of men were recommended for tenure in the years surrounding the Namenwirth decision. Perhaps the most pertinent case is that of Dr. Timothy Moermond, who was recommended for tenure in the 1978-79 academic year. Moermond’s publication record was similar to Namenwirth’s; like hers, it seemed to improve as the time for the tenure decision approached. Among his letters from outside evaluators, some were positive and some were critical. At the time the department first considered him for tenure, he had no outside grant support. The vote in the department was unanimous in favor of tenure, and he was recommended to the Division Executive Committee.
Namenwirth relies in part on the difference between the way she was treated and the way Moermond was treated to support her claim of discrimination. She points to these particular differences:
—The rate of publication for each of them increased toward tenure time. In the department’s recommendation for Moermond, this was described as an expression of his meticulous approach to research. In the Namenwirth packet, it was described as the result of a last minute effort.
—The supervision of graduate students is an important part of teaching. The number of graduate students reported as being supervised by Namenwirth was lower than the actual number; according to Namenwirth, the number reported for Moermond was inflated.
—Each received critical comments from outside evaluators. Negative comments in the Moermond letters were minimized in the summary in his tenure packet sent to the Division. Negative comments about Namenwirth were highlighted.2
In spite of the unanimous recommendation from the Department, the Division Executive Committee found Moermond’s record to be weak, and it voted to deny him tenure. Three Zoology faculty members appeared before the committee to argue his case. They presented additional material in his favor, including evidence of a grant he had received after the tenure packet had first been sent on. The Executive Committee reconsidered and recommended Moermond for tenure. Shortly thereafter he was promoted to a tenured rank. The Department had made no such effort in the Namenwirth case.
On July 25, 1979, Namenwirth filed a complaint with the Milwaukee office of the Federal Contract Compliance Division of the United States Department of Labor, alleging sex discrimination. After an investigation, the Department of Labor concluded that she had been denied tenure because of her sex. This determination was based, among other things, on the poor record of the University in having women in tenured positions and on the failure of the Department to advocate tenure for Namenwirth as aggressively as for Moermond.
*1240In October, 1983, Namenwirth brought an action in the District Court for the Western District of Wisconsin, claiming a deprivation of rights secured by Title VII of the Civil Rights Act of 1964, as amended by the Equal Opportunity Act of 1972, 42 U.S.C. § 2000e, et seq. She alleged that the University denial of tenure was the result of disparate treatment. With consent of the parties, the case was brought before a magistrate under 28 U.S.C. § 636(c)(1). The magistrate found that Namenwirth had discharged her burden of establishing a prima facie case of disparate treatment because of sex; that the University had rebutted that case by arguing that her record did not satisfy the standard of excellence required by the University of Wisconsin at Madison; that the University’s argument was genuine and reasonable; and that Namenwirth therefore failed to carry her ultimate burden of persuasion. He held that the denial of tenure was not based upon her sex, and dismissed her complaint with prejudice. On appeal, Namenwirth argues that the lower court misapplied the law; that it was not the duty of the magistrate to inquire into the reasonableness of the motive offered by the University, but rather to examine the evidence that the motive was pretextual.
II.
In a disparate treatment suit the plaintiff is required to establish a prima facie case of discrimination; if she succeeds, the university is called upon to give a non-discriminatory reason for failing to give her tenure; and, when the university has come forward with such a reason, the burden then rests with the plaintiff to prove discrimination. McDonnell Douglas v. Green, 411 U.S. 792, 807, 93 S.Ct. 1817, 1826, 36 L.Ed.2d 668 (1973). To establish a prima facie ease of discrimination in tenure, the plaintiff must show that: (1) she is a member of a protected class; (2) she was qualified for tenure; (3) she was denied tenure; and (4) an applicant not in the protected class was granted tenure. Cf Smith v. University of North Carolina, 632 F.2d 316, 332 (4th Cir.1980). She may attack the motive advanced by the university in one of two ways: she may try to show that the motive is not worthy of belief; or she may try to show that in fact some other motive explains the university’s action better than the motive offered. Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 256, 101 S.Ct. 1029, 1095, 67 L.Ed.2d 207 (1981). In either case she is not called upon to offer direct proof of discriminatory motive. International Brotherhood of Teamsters v. United States, 431 U.S. 324, 358 n. 44, 97 S.Ct. 1843, 1866, 52 L.Ed.2d 396 (1977).
To prove the proffered motive is not worthy of belief, evidence of a comparative sort is appropriate: if others were hired or promoted though by the same reasoning they ought to have been excluded, then the motive is a “pretext.” Mozee v. Jeffboat, 746 F.2d 365, 369 (7th Cir.1984). It is not determinative that the motive as applied to the plaintiff is “genuine,” or “reasonable;” if it does not apply to all equally, it does not rebut a charge of discrimination. Thus, suppose a rule forbids the wearing of bracelets near certain machinery. Such a rule may indeed be reasonable, and the fact that she wore a bracelet near the machinery may be the reason — a genuine reason— that a certain woman was fired. But if men are exempted from the rule, whether intentionally or not, so that men who wear bracelets are never fired, then the reason is “pretext” in the technical sense given to that word by the law. Obviously comparative evidence is relevant in determining whether a motive is pretextual. As with the bracelets, the reason in isolation may be both genuine and reasonable, but that is not enough to show that its effect is not discriminatory. To show that, it must be shown that it does not work to the detriment of the protected class alone.
The state of Wisconsin has argued in its brief to this court that comparisons *1241are out of place in a tenure decision. We disagree. Comparisons may be more difficult in the case of professional and academic employment decisions, but they may be essential to a determination of discrimination; and where they are, and where the evidence is available, they must be made.
III.
The magistrate rejected many of the findings of fact proposed by the parties. The trier of fact has a substantial amount of discretion in this matter. Where it is clear that he has considered the evidence, we will not remand simply to have him produce more detailed findings of fact.
In this case it is true that the magistrate might have made more detailed findings, but it is also clear that he considered and took account of the relevant evidence. He elected not to make many of the findings suggested, many of which dealt with minutiae. He recognized and considered the issue of prior findings of discrimination by the University; the discriminatory treatment of the single woman to have been tenured by the department; and the comparative evidence concerning the qualification of males tenured at about the time Namenwirth was denied tenure. Since we believe that the magistrate did take account of the evidence, we do not believe that more extensive and detailed findings would change the outcome.
The plaintiff claims that the magistrate applied an incorrect standard. The magistrate found that the motive offered by the University for not offering Namenwirth tenure, that the quantity of her research output and the potential of her projected research did not satisfy the University’s standard of excellence, was a reasonable one:
I am persuaded that the defendant Board’s articulated reason for denying tenure was the true and genuine basis for the challenged denial of tenure: a reasonable decision, with which others equipped to render the decision might differ, but which I cannot reject as arbitrary, that Namenwirth did not possess the requisite research credentials to satisfy the tenure standard.
It is true that if the magistrate had addressed himself to the University’s proffered motive as a motive in the Namenwirth case only, isolated from other tenure determinations, then he would not have applied the correct standard. As we have already pointed out, to find merely that the motive was genuine (that is, that it really was the motive in this case) and reasonable does not suffice to show that it was not pretextual, in the sense required by law. For if the same explanation, however genuine or reasonable, were used to exclude only members of a protected class — as when, in the example above, women were fired for the wearing of bracelets but men were not — then in the eyes of the law the explanation would be pretextual. So here, if the proffered motive, however genuine in the particular case, however reasonable in general, operated only to exclude women, it would be pretextual. And if the magistrate considered only the genuineness and reasonableness of the motive in isolation, he did not apply the correct standard.
But the magistrate did apply the correct standard in this case. He might have worded his conclusion a little more felicitously. He might, for example, have been more specific, finding that in the eyes of the department Moermond showed greater promise than Namenwirth. But he did consider the comparative evidence; he did explicitly find that men in Namenwirth’s position would not have been treated differently,3 and he concluded that the *1242motive was not “arbitrary.” A motive which is applied in some cases and not in others, not along lines of principle (or of acceptable principle), is arbitrary, and we cannot imagine what the magistrate would have meant by the use of the word if he did not mean that the University’s standards of excellence were applied to all equally.
Since we find that the magistrate has adequately considered the evidence, and has applied the appropriate standard, it remains for us to determine whether his finding of no discrimination is clearly erroneous.
IV.
Namenwirth argues that she is qualified and ought to have been tenured. The magistrate found in fact that she was qualified; that was the basis on which he found that she had made a prima facie case. But it does not follow that she ought to have been awarded tenure-. Mere qualification depends on objective measures— the terminal degree, the number of publications, and so on. Tenure requires something more; it requires that the department believe that the candidate have a certain amount of promise. The -magistrate compared her record with the record of men who were awarded tenure and determined that the claim that the Department — and the University — found insufficient promise in Namenwirth’s work was not pretextual. We have examined the records of Namenwirth and the various male candidates, and we cannot say that the magistrate’s conclusion is in clear conflict with the evidence.
The evidence consists for the most part of comparative evidence of research quality and potential. That means that it consists in large part of the opinions of academics who serve not only as experts on qualifications but also as decision-makers in the tenure process. It is not our role, as federal courts have acknowledged, to consider merely the hard evidence of research output and hours spent on committee work, and reach tenure determinations de novo. A crucial part of the evidence we rely on is the esteem in which the candidate is held by the very persons making the tenure decision.
Thus, in the case before us, Namenwirth’s bid for tenure was defeated primarily by the close departmental vote and the equivocal recommendation which issued from that vote. If we consider the department vote as an assessment of qualification by experts in the field — and there is no doubt that it is that — then the University's denial of tenure is based on rather hard evidence that Namenwirth was not as well qualified as those who were granted tenure. But the department vote is not just an independent assessment of qualifications; it is actually the first step in the decision process, and for most candidates it is the most important step. For many candidates it is the department that actually makes the decisions, and it is clear that it was the department that made the decision in Namenwirth’s case.
*1243To allow the decision-maker also to act as the source of judgments of qualification would ordinarily defeat the purpose of the discrimination laws. But in the case of tenure decisions we see no alternative. Tenure decisions have always relied primarily on judgments about academic potential, and there is no algorithm for producing those judgments. Given the similar research output of two candidates, an experienced faculty committee might — quite rightly — come to different conclusions about the potential of the candidates. It is not our place to question the significance or validity of such conclusions.
But to say all that is only to face up to the problem. The problem remains: faculty votes should not be permitted to camouflage discrimination, even the unconscious discrimination of well-meaning and established scholars. The courts have struggled with the problem since Title VII was extended to the university, and have found no solution. Because of the way we have described the problem — the decision-maker is also the source of the qualifications— there may be no solution; winning the esteem of one’s colleagues is just an essential part of securing tenure. And that seems to mean that in a case of this sort, where it is a matter of comparing qualification against qualification, the plaintiff is bound to lose.
But there are other sorts of cases. There are cases that involve a pattern of discrimination. Hazelwood School District v. United States, 433 U.S. 299, 307, 97 S.Ct. 2736, 2741, 53 L.Ed.2d 768 (1977). There are cases in which procedural barriers are placed in the way of members of a certain class. See, e.g., Kunda v. Muhlenberg College, 621 F.2d 532, 546 (3d Cir.1980). There are cases that involve outright discriminatory judgments. See, e.g., Lynn v. Regents of the University of California, 656 F.2d 1337, 1343 (9th Cir.1981), cert. denied, 459 U.S. 823, 103 S.Ct. 53, 74 L.Ed.2d 59 (1982). Thus, although we may despair of extricating discriminatory motives from collegial judgments about potential and worth, the outlook is not entirely bleak.
In this case, there was some evidence of past discrimination on the part of both the University and the Department. That evidence could not offset the rather substantial evidence of the close department vote; and while such a vote might itself be an indication of bias, on the evidence before him the magistrate was not clearly wrong in deciding that in this case no bias was involved.
While Namenwirth’s record was in many ways similar to the records of men who were tenured in the same time period, there were also differences which would have justified the difference in the votes. Moermond, for example, was the editor of a respected professional journal. We have said that we would not review the evidence de novo, and we will not; but the clear drift of the evidence is this: that the Department anguished over the Ñamenwirth decision and that in the end its equivocal recommendation to the Division reflected its own uncertainty; that it felt varying degrees of uncertainty about the male candidates, but in no case as serious an uncertainty as in Namenwirth’s case; and that its unwillingness to press Namenwirth’s case as strongly as it pressed the cases of the others was based on perceptions of promise.
We conclude that the magistrate’s finding of no discrimination was supported by the evidence, and that in this case there was no convincing evidence of other sorts to counter that evidence.
The judgment of the district court is therefore affirmed.
. The magistrate found that the standards for tenure in the Department did not change between 1968 and 1979.
. At the trial before the magistrate, Namenwirth presented the testimony of Judith Laws, a social psychologist who has done research on sex bias in the workplace. Dr. Laws testified that such differences in treatment are indications of the presence of sexual bias.
. The magistrate’s last three findings of fact were the following:
81. In the course of Namenwirth’s tenure consideration, a sizable minority of the Zoology Department tenured faculty in fact decided that Namenwirth did not meet the standard for tenure at the UW-Madison, principally because of the deficiencies they perceived in her research output and potential. This was a reasonable decision, and the articulated rea*1242son for the decision is not a pretext. If all other factors had been the same, but Namenwirth had been a man and not a woman, the decision of the sizable minority of the Zoology Department would have been the same: to oppose Namenwirth's tenure.
82. In April of 1977 and again in October of 1978, the Divisional Committee in fact decided that Namenwirth did not meet the standard for tenure at the UW-Madison, particularly in the area of research. This was a reasonable decision, and the articulated reason for the decision was not a pretext. If all other factors had been the same, but Namenwirth had been a man and not a woman, the decision of the Divisional Committee would have been the same: to reject Namenwirth’s tenure.
83. In May of 1977 and again in February of 1979, Dean David Cronon in fact decided that Namenwirth did not meet the standard for tenure at UW-Madison, particularly in the area of research. This was a reasonable decision, and the articulated reason for it was not a pretext. If all other factors had been the same, and if Namenwirth had been a man and not a woman, the decision of Dean Cronon would have been the same: to deny tenure to Namenwirth.