with whom The Chief Justice, Justice Rehnquist, and Justice O’Connor join, dissenting.
In past decisions, this Court has been sensitive to the critical difference between cases proving discrimination under Title VII, 42 U. S. C. §2000e et seq. (1976 ed. and Supp. IV), by a showing of disparate treatment or discriminatory intent and those proving such discrimination by a showing of disparate impact. Because today’s decision blurs that distinction and results in a holding inconsistent with the very nature of disparate-impact claims, I dissent.
hH
Section 703(a)(2) of Title VII, 42 U. S. C. §.2000e-2(a)(2), provides that it is an unlawful employment practice for an employer to
“limit, segregate, or classify his employees or applicants for employment in any way which would deprive or tend to deprive any individual of employment opportunities or *457otherwise adversely affect his status as an employee, because of such individual’s race, color, religion, sex, or national origin.”
Although this language suggests that discrimination occurs only on an individual basis, in Griggs v. Duke Power Co., 401 U. S. 424, 432 (1971), the Court held that discriminatory intent on the part of the employer against an individual need not be shown when “employment procedures or testing mechanisms . . . operate as ‘built-in headwinds’ for minority groups and are unrelated to measuring job capability.” Thus, the Court held that the “disparate impact” of an employer’s practices on a racial group can violate § 703(a)(2) of Title VII. In Griggs and each subsequent disparate-impact case, however, the Court has considered, not whether the claimant as an individual had been classified in a manner impermissible under § 703(a)(2), but whether an employer’s procedures have had an adverse impact on the protected group to which the individual belongs.
Thus, while disparate-treatment cases focus on the way in which an individual has been treated, disparate-impact cases are concerned with the protected group. This key distinction was explained in Furnco Construction Corp. v. Waters, 438 U. S. 567, 581-582 (1978) (Marshall, J., concurring in part):
“It is well established under Title VII that claims of employment discrimination because of race may arise in two different ways. Teamsters v. United States, 431 U. S. 324, 335-336, n. 15 (1977). An individual may allege that he has been subjected to ‘disparate treatment’ because of his race, or that he has been the victim of a facially neutral practice having a ‘disparate impact’ on his racial group.”1
*458In keeping with this distinction, our disparate-impact cases consistently have considered whether the result of an employer’s total selection process had an adverse impact upon the protected group.2 If this case were decided by reference to the total process — as our cases suggest that it should be— the result would be clear. Here 22.9% of the blacks who entered the selection process were ultimately promoted, compared with only 13.5% of the whites. To say that this selection process had an unfavorable “disparate impact” on blacks is to ignore reality.
The Court, disregarding the distinction drawn by our cases, repeatedly asserts that Title VII was designed to protect individual, not group, rights. It emphasizes that some individual blacks were eliminated by the disparate impact of the preliminary test. But this argument confuses the aim of Title VII with the legal theories through which its aims were intended to be vindicated. It is true that the aim of Title VII is to protect individuals, not groups. But in advancing this commendable objective, Title VII jurisprudence has recognized two distinct methods of proof. In one set of cases— those involving direct proof of discriminatory intent — the plaintiff seeks to establish direct, intentional discrimination against him. In that type of case, the individual is at the forefront throughout the entire presentation of evidence. In disparate-impact cases, by contrast, the plaintiff seeks to carry his burden of proof by way of inference — by showing that an employer’s selection process results in the rejection of a disproportionate number of members of a protected group *459to which he belongs. From such a showing a fair inference then may be drawn that the rejected applicant, as a member of that disproportionately excluded group, was himself a victim of that process’ ‘“built-in headwinds.’” Griggs, supra, at 432. But this method of proof — which actually defines disparate-impact theory under Title VII — invites the plaintiff to prove discrimination by reference to the group rather than to the allegedly affected individual.3 There can be no violation of Title VII on the basis of disparate impact in the absence of disparate impact on a group.4
In this case respondent black employees seek to benefit from a conflation of “discriminatory treatment” and “disparate impact” theories. But they cannot have it both ways. Having undertaken to prove discrimination by reference to one set of group figures (used at a preliminary point in the selection process), these respondents then claim that wowdis-crimination cannot be proved by viewing the impact of the entire process on the group as a whole. The fallacy of this reasoning — accepted by the Court — is transparent. It is to *460confuse the individualistic aim of Title VII with the methods of proof by which Title VII rights may be vindicated. The respondents, as individuals, are entitled to the full personal protection of Title VII. But, having undertaken to prove a violation of their rights by reference to group figures, respondents cannot deny petitioners the opportunity to rebut their evidence by introducing figures of the same kind. Having pleaded a disparate-impact case, the plaintiff cannot deny the defendant the opportunity to show that there was no disparate impact. As the Court of Appeals for the Third Circuit noted in EEOC v. Greyhound Lines, Inc., 635 F. 2d 188, 192 (1980);
“[N]o violation of Title VII can be grounded on the disparate impact theory without proof that the questioned policy or practice has had a disproportionate impact on the employer’s workforce. This conclusion should be as obvious as it is tautological: there can be no disparate impact unless there is [an ultimate] disparate impact.”
Where, under a facially neutral employment process, there has been no adverse effect on the group — and certainly there has been none here — Title VII has not been infringed.
I — I HH
The Court’s position is no stronger in case authority than it is in logic. None of the cases relied upon by the Court controls the outcome of this case.5 Indeed, the disparate-*461impact cases do not even support the propositions for which they are cited. For example, the Court cites Dothard v. Rawlinson, 433 U. S. 321 (1977) (holding impermissible minimum statutory height and weight requirements for correctional counselors), and observes that “[although we noted in passing that women constituted 36.89 percent of the labor force and only 12.9 percent of correctional counselor positions, our focus was not on this ‘bottom line.’ We focused instead on the disparate effect that the minimum height and weight standards had on applicants: classifying far more women than men as ineligible for employment.” Ante, at 450. In Dothard, however, the Court was not considering a case in which there was any difference between the discriminatory effect of the employment standard and the number of minority members actually hired. The Dothard Court itself stated:
“[T]o establish a prima facie case of discrimination, a plaintiff need only show that the facially neutral standards in question select applicants for hire in a discriminatory pattern. Once it is shown that the employment standards are discriminatory in effect, the employer must meet ‘the burden of showing that any given requirement [has] ... a manifest relationship to the employment in question.’” 433 U. S., at 329 (emphasis added).
The Dothard Court did not decide today’s case. It addressed only a case in which the challenged standards had a discriminatory impact at the bottom line — the hiring decision. And the Dothard Court’s “focus,” referred to by the Court, is of no help in deciding the instant case.6
*462The Court concedes that the other major cases on which it relies, Furnco, Los Angeles Dept. of Water & Power v. Manhart, 435 U. S. 702 (1978), and Phillips v. Martin Marietta Corp., 400 U. S. 542 (1971) (per curiam) “involved facially discriminatory policies, while the claim in the instant case is one of discrimination from a facially neutral policy.” Ante, at 455. The Court nevertheless applies the principles derived from those cases to the case at bar. It does so by reiterating the view that Title VII protects individuals, not groups, and therefore that the manner in which an employer has treated other members of a group cannot defeat the claim of an individual who has suffered as a result of even a facially neutral policy. As appealing as this sounds, it confuses the distinction — uniformly recognized until today — between disparate impact and disparate treatment. See supra, at 457-458. Our cases, cited above, have made clear that discriminatory-impact claims cannot be based on how an individual is treated in isolation from the treatment of other members of the group. Such claims necessarily are based on whether the group fares less well than other groups under a policy, practice, or test. Indeed, if only one minority member has *463taken a test, a disparate-impact claim cannot be made, regardless of whether the test is an initial step in the selection process or one of several factors considered by the employer in making an employment decision.7
Ill
Today’s decision takes a long and unhappy step in the direction of confusion. Title VII does not require that employers adopt merit hiring or the procedures most likely to permit the greatest number of minority members to be considered for or to qualify for jobs and promotions. See Texas Dept. of Community Affairs v. Burdine, 450 U. S. 248, 258-259 (1981); Furnco, 438 U. S., at 578. Employers need not develop tests that accurately reflect the skills of every individual candidate; there are few if any tests that do so. Yet the Court seems unaware of this practical reality, and perhaps oblivious to the likely consequences of its decision. By its holding today, the Court may force employers either to eliminate tests or rely on expensive, job-related, testing procedures, the validity of which may or may not be sustained if challenged. For state and local governmental employers with limited funds, the practical effect of today’s decision may well be the adoption of simple quota hiring.8 This arbi*464trary method of employment is itself unfair to individual applicants, whether or not they are members of minority groups. And it is not likely to produce a competent work force. Moreover, the Court’s decision actually may result in employers employing fewer minority members. As Judge Newman noted in Brown v. New Haven Civil Service Board, 474 F. Supp. 1256, 1263 (Conn. 1979):
“[A]s private parties are permitted under Title VII itself to adopt voluntary affirmative action plans, . . . Title VII should not be construed to prohibit a municipality’s using a hiring process that results in a percentage of minority policemen approximating their percentage of the local population, instead of relying on the expectation that a validated job-related testing procedure will produce an equivalent result, yet with the risk that it might lead to substantially less minority hiring.”
Finding today’s decision unfortunate in both its analytical approach and its likely consequences, I dissent.
See also Teamsters v. United States, 431 U. S. 324, 335-336, n. 15 (1977) (similar explanation).
See Dothard v. Rawlinson, 433 U. S. 321, 329 (1977) (statutory height and weight requirements operated as a bar to employment of disproportionate number of women); Albemarle Paper Co. v. Moody, 422 U. S. 405, 409-411 (1975) (seniority system allegedly locked blacks into lower paying jobs; applicants to skilled lines of progression were required to pass two tests); Griggs v. Duke Power Co., 401 U. S. 424, 431 (1971) (tests were an absolute bar to transfers or hiring; the Court observed that all Congress requires is “the removal of artificial, arbitrary, and unnecessary barriers to employment. . .”) (emphasis added).
Initially, the plaintiff bears the burden of establishing a prima facie case that Title VII has been infringed. See Texas Dept. of Community Affairs v. Burdine, 450 U. S. 248, 252-253 (1981). In a disparate-impact case, this burden is met by showing that an employer’s selection process results in the rejection of a disproportionate number of members of a protected group. See Teamsters v. United States, supra, at 336-338. Regardless of whether the plaintiff’s prima facie case must itself focus on the defendant’s overall selection process or whether it is sufficient that the plaintiff establish that at least one pass-fail barrier has resulted in disparate impact, the employer’s presentation of evidence showing that its overall selection procedure does not operate in a discriminatory fashion certainly dispels any inference of discrimination. In such instances, at the close of the evidence, the plaintiff has failed to show disparate impact by a preponderance of the evidence.
The Equal Employment Opportunity Commission and other federal enforcement agencies have adopted the “bottom-line” principle — i. e., the process viewed as a whole — in deciding when to bring an action against an employer. See Uniform Guidelines on Employee Selection Procedures, 5 CFR § 300.103(c) (1981).
The Court concentrates on cases of questionable relevance. Most of the lower courts that have squarely considered the question have concluded that there can be no violation of Title VII on a disparate-impact basis when there is no disparate impact at the bottom line. See, e. g., EEOC v. Greyhound Lines, Inc., 635 F. 2d 188 (CA3 1980); EEOC v. Navajo Refining Co., 593 F. 2d 988 (CA10 1979); Friend v. Leidinger, 588 F. 2d 61, 66 (CA4 1978); Rule v. International Assn. of Ironworkers, 568 F. 2d 558 (CA8 1977); Smith v. Troyan, 520 F. 2d 492,497-498 (CA6 1975), cert. denied, 426 U. S. 934 (1976); Williams v. City & County of San Francisco, 483 F. Supp. 335 (ND Cal. 1979); Brown v. New Haven Civil Service *461Board, 474 F. Supp. 1256 (Conn. 1979); Lee v. City of Richmond, 456 F. Supp. 756 (ED Va. 1978).
The Court cites language from two other disparate-impact cases. The Court notes that in Albemarle Paper Co. v. Moody, 422 U. S. 405 (1975), the Court “remanded to allow the employer to attempt to show that the tests . . . given ... for promotion were job related.” Ante, at 450. But the fact that the Court did so without suggesting “that by promoting a suf*462ficient number of black employees who passed the examination, the employer could avoid this burden,” ibid., can hardly be precedent for the negative of that proposition when the issue was neither presented in the facts of the case nor addressed by the Court.
Similarly, New York Transit Authority v. Beazer, 440 U. S. 568 (1979), provides little support despite the language quoted by the Court. See ante, at 450, quoting 440 U. S., at 584 (“ ‘A prima facie violation of the Act may be established by statistical evidence showing that an employment practice has the effect of denying members of one race equal access to employment opportunities’”) (emphasis added by the Court). In Beazer, the Court ruled that the statistical evidence actually presented was insufficient to establish a prima facie case of discrimination, and in doing so it indicated that it would have found statistical evidence of the number of applicants and employees in a methadone program quite probative. See id., at 585. Beazer therefore does not justify the Court’s speculation that the number of blacks and Hispanics actually employed were irrelevant to whether a case of disparate impact had been established under Title VII.
Courts have recognized that the probative value of statistical evidence varies with sample size in disparate-impact cases. See, e. g., Teamsters v. United States, 431 U. S., at 340, n. 20 (“Considerations such as small sample size may, of course, detract from the value of such evidence . . Mayor of Philadelphia v. Educational Equality League, 415 U. S. 605, 621 (1974) (“[T]he District Court’s concern for the smallness of the sample presented by the 13-member Panel was . . . well founded”); Rogillio v. Diamond Shamrock Chemical Co., 446 F. Supp. 423, 427-428 (SD Tex. 1978) (sample of 10 too small); Dendy v. Washington Hospital Center, 431 F. Supp. 873, 876 (DC 1977) (sample must be “large enough to mirror the reality of the employment situation”). A sample of only one would have far too little probative value to establish a prima facie case of disparate impact.
Another possibility is that employers may integrate consideration of test results into one overall hiring decision based on that “factor” and addi*464tional factors. Such a process would not, even under the Court’s reasoning, result in a finding of discrimination on the basis of disparate impact unless the actual hiring decisions had a disparate impact on the minority group. But if employers integrate test results into a single-step decision, they will be free to select only the number of minority candidates proportional to their representation in the work force. If petitioners had used this approach, they would have been able to hire substantially fewer blacks without liability on the basis of disparate impact. The Court hardly could have intended to encourage this.