Virginia Military Institute (VMI) is the sole single-sex school among Virginia's public institutions of higher learning. VMI's distinctive mission is to produce "citizen-soldiers," men prepared for leadership in civilian life and in military service. . . . The United States sued Virginia and VMI, alleging that VMI's exclusively male admission policy violated the Fourteenth Amendment's Equal Protection Clause. . . . In response, Virginia proposed a parallel program for women: Virginia Women's Institute for Leadership (VWIL), located at Mary Baldwin College, a private liberal arts school for women. The District Court found that Virginia's proposal satisfied the Constitution's equal protection requirement, and the Fourth Circuit affirmed. The appeals court deferentially reviewed Virginia's plan and determined that provision of single-gender educational options was a legitimate objective. Maintenance of single-sex programs, the court concluded, was essential to that objective. The court recognized, however, that its analysis risked bypassing equal protection scrutiny, so it fashioned an additional test, asking whether VMI and VWIL students would receive ``substantively comparable'' benefits. Although the Court of Appeals acknowledged that the VWIL degree lacked the historical benefit and prestige of a VMI degree, the court nevertheless found the educational opportunities at the two schools sufficiently comparable.

 Ginsburg, J., delivered the opinion of the Court, in which Stevens, O'Connor, Kennedy, Souter, and Breyer, JJ., joined. Rehnquist, C. J., filed an opinion concurring in the judgment. Scalia, J., filed a dissenting opinion. Thomas, J., took no part in the consideration or decision of the case.

Justice Ginsburg delivered the opinion of the Court.

 Virginia's public institutions of higher learning include an incomparable military college, Virginia Military Institute (VMI). The United States maintains that the Constitution's equal protection guarantee precludes Virginia from reserving exclusively to men the unique educational opportunities VMI affords. We agree.

 VMI has notably succeeded in its mission to produce leaders; among its alumni are military generals, Members of Congress, and business executives. The school's alumni overwhelmingly perceive that their VMI training helped them to realize their personal goals. VMI's endowment reflects the loyalty of its graduates; VMI has the largest per-student endowment of all undergraduate institutions in the Nation.

 Neither the goal of producing citizen-soldiers nor VMI's implementing methodology is inherently unsuitable to women. And the school's impressive record in producing leaders has made admission desirable to some women. Nevertheless, Virginia has elected to preserve exclusively for men the advantages and opportunities a VMI education affords.

 VMI today enrolls about 1,300 men as cadets.2 Its academic offerings in the liberal arts, sciences, and engineering are also available at other public colleges and universities in Virginia. But VMI's mission is special. It is the mission of the school

"`to produce educated and honorable men, prepared for the varied work of civil life, imbued with love of learning, confident in the functions and attitudes of leadership, possessing a high sense of public service, advocates of the American democracy and free enterprise system, and ready as citizen-soldiers to defend their country in time of national peril.'"

 VMI produces its "citizen-soldiers" through "an adversative, or doubting, model of education" which features "[p]hysical rigor, mental stress, absolute equality of treatment, absence of privacy, minute regulation of behavior, and indoctrination in desirable values." As one Commandant of Cadets described it, the adversative method "dissects the young student," and makes him aware of his "limits and capabilities," so that he knows "how far he can go with his anger, . . . how much he can take under stress, . . . exactly what he can do when he is physically exhausted."

 VMI cadets live in spartan barracks where surveillance is constant and privacy nonexistent; they wear uniforms, eat together in the mess hall, and regularly participate in drills. Entering students are incessantly exposed to the rat line, "an extreme form of the adversative model," comparable in intensity to Marine Corps boot camp.  Tormenting and punishing, the rat line bonds new cadets to their fellow sufferers and, when they have completed the 7-month experience, to their former tormentors.

 VMI's "adversative model" is further characterized by a hierarchical "class system" of privileges and responsibilities, a "dyke system" for assigning a senior class mentor to each entering class "rat," and a stringently enforced "honor code," which prescribes that a cadet "`does not lie, cheat, steal nor tolerate those who do.'"

 VMI attracts some applicants because of its reputation as an extraordinarily challenging military school, and "because its alumni are exceptionally close to the school." "[W]omen have no opportunity anywhere to gain the benefits of [the system of education at VMI]."

 In the two years preceding the lawsuit, the District Court noted, VMI had received inquiries from 347 women, but had responded to none of them. "[S]ome women, at least," the court said, "would want to attend the school if they had the opportunity." The court further recognized that, with recruitment, VMI could "achieve at least 10% female enrollment"-"a sufficient `critical mass' to provide the female cadets with a positive educational experience." And it was also established that "some women are capable of all of the individual activities required of VMI cadets." In addition, experts agreed that if VMI admitted women, "the VMI ROTC experience would become a better training program from the perspective of the armed forces, because it would provide training in dealing with a mixed-gender army."

 The District Court ruled in favor of VMI, however, and rejected the equal protection challenge pressed by the United States. That court correctly recognized that Mississippi Univ. for Women v. Hogan, 458 U.S. 718 (1982), was the closest guide. There, this Court underscored that a party seeking to uphold government action based on sex must establish an "exceedingly persuasive justification" for the classification. To succeed, the defender of the challenged action must show "at least that the classification serves important governmental objectives and that the discriminatory means employed are substantially related to the achievement of those objectives."

 The District Court reasoned that education in "a single-gender environment, be it male or female," yields substantial benefits. VMI's school for men brought diversity to an otherwise coeducational Virginia system, and that diversity was "enhanced by VMI's unique method of instruction." If single-gender education for males ranks as an important governmental objective, it becomes obvious, the District Court concluded, that the only means of achieving the objective "is to exclude women from the all-male institution-VMI."

"Women are [indeed] denied a unique educational opportunity that is available only at VMI," the District Court acknowledged. But "[VMI's] single-sex status would be lost, and some aspects of the [school's] distinctive method would be altered" if women were admitted: "Allowance for personal privacy would have to be made,"; "[p]hysical education requirements would have to be altered, at least for the women,"; the adversative environment could not survive unmodified. Thus, "sufficient constitutional justification" had been shown, the District Court held, "for continuing [VMI's] single-sex policy."

 The appeals court greeted with skepticism Virginia's assertion that it offers single-sex education at VMI as a facet of the State's overarching and undisputed policy to advance "autonomy and diversity." The court underscored Virginia's nondiscrimination commitment: "`[I]t is extremely important that [colleges and universities] deal with faculty, staff, and students without regard to sex, race, or ethnic origin.'" "That statement," the Court of Appeals said, "is the only explicit one that we have found in the record in which the Commonwealth has expressed itself with respect to gender distinctions." Furthermore, the appeals court observed, in urging "diversity" to justify an all-male VMI, the State had supplied "no explanation for the movement away from [single-sex education] in Virginia by public colleges and universities." In short, the court concluded, "[a] policy of diversity which aims to provide an array of educational opportunities, including single-gender institutions, must do more than favor one gender."

 The parties agreed that "some women can meet the physical standards now imposed on men," and the court was satisfied that "neither the goal of producing citizen soldiers nor VMI's implementing methodology is inherently unsuitable to women". The Court of Appeals, however, accepted the District Court's finding that "at least these three aspects of VMI's program-physical training, the absence of privacy, and the adversative approach-would be materially affected by coeducation." Remanding the case, the appeals court assigned to Virginia, in the first instance, responsibility for selecting a remedial course. The court suggested these options for the State: Admit women to VMI; establish parallel institutions or programs; or abandon state support, leaving VMI free to pursue its policies as a private institution. . . .
 

 The average combined SAT score of entrants at Mary Baldwin is about 100 points lower than the score for VMI freshmen. Mary Baldwin's faculty holds "significantly fewer Ph.D.'s than the faculty at VMI," and receives significantly lower salaries. While VMI offers degrees in liberal arts, the sciences, and engineering, Mary Baldwin, at the time of trial, offered only bachelor of arts degrees. A VWIL student seeking to earn an engineering degree could gain one, without public support, by attending Washington University in St. Louis, Missouri, for two years, paying the required private tuition.

 Experts in educating women at the college level composed the Task Force charged with designing the VWIL program; Task Force members were drawn from Mary Baldwin's own faculty and staff. Training its attention on methods of instruction appropriate for "most women," the Task Force determined that a military model would be "wholly inappropriate" for VWIL.

 VWIL students would participate in ROTC programs and a newly established, "largely ceremonial" Virginia Corps of Cadets, but the VWIL House would not have a military format, and VWIL would not require its students to eat meals together or to wear uniforms during the school day. In lieu of VMI's adversative method, the VWIL Task Force favored "a cooperative method which reinforces self-esteem." In addition to the standard bachelor of arts program offered at Mary Baldwin, VWIL students would take courses in leadership, complete an off-campus leadership externship, participate in community service projects, and assist in arranging a speaker series.

 Virginia represented that it will provide equal financial support for in-state VWIL students and VMI cadets, and the VMI Foundation agreed to supply a $5.4625 million endowment for the VWIL program, id., at 499. Mary Baldwin's own endowment is about $19 million; VMI's is $131 million. Mary Baldwin will add $35 million to its endowment based on future commitments; VMI will add $220 million. The VMI Alumni Association has developed a network of employers interested in hiring VMI graduates. The Association has agreed to open its network to VWIL graduates, but those graduates will not have the advantage afforded by a VMI degree.

 A divided Court of Appeals affirmed the District Court's judgment. This time, the appellate court determined to give "greater scrutiny to the selection of means than to the [State's] proffered objective." The official objective or purpose, the court said, should be reviewed deferentially. Respect for the "legislative will," the court reasoned, meant that the judiciary should take a "cautious approach," inquiring into the "legitima[cy]" of the governmental objective and refusing approval for any purpose revealed to be "pernicious."

"[P]roviding the option of a single-gender college education may be considered a legitimate and important aspect of a public system of higher education," the appeals court observed; that objective, the court added, is "not pernicious,". Moreover, the court continued, the adversative method vital to a VMI education "has never been tolerated in a sexually heterogeneous environment." The method itself "was not designed to exclude women," the court noted, but women could not be accommodated in the VMI program, the court believed, for female participation in VMI's adversative training "would destroy . . . any sense of decency that still permeates the relationship between the sexes."

 The court recognized that, as it analyzed the case, means merged into end, and the merger risked "bypass[ing] any equal protection scrutiny." The court therefore added another inquiry, a decisive test it called "substantive comparability." The key question, the court said, was whether men at VMI and women at VWIL would obtain "substantively comparable benefits at their institution or through other means offered by the [S]tate." Although the appeals court recognized that the VWIL degree "lacks the historical benefit and prestige" of a VMI degree, it nevertheless found the educational opportunities at the two schools "sufficiently comparable."

 . . .
 

 Today's skeptical scrutiny of official action denying rights or opportunities based on sex responds to volumes of history. As a plurality of this Court acknowledged a generation ago, "our Nation has had a long and unfortunate history of sex discrimination." Frontiero v. Richardson, 411 U.S. 677, 684 (1973). Through a century plus three decades and more of that history, women did not count among voters composing "We the People";5 not until 1920 did women gain a constitutional right to the franchise. And for a half century thereafter, it remained the prevailing doctrine that government, both federal and state, could withhold from women opportunities accorded men so long as any "basis in reason" could be conceived for the discrimination. . . .

 In 1971, for the first time in our Nation's history, this Court ruled in favor of a woman who complained that her State had denied her the equal protection of its laws. Reed v. Reed, 404 U.S. 71, 73. . . . Since Reed, the Court has repeatedly recognized that neither federal nor state government acts compatibly with the equal protection principle when a law or official policy denies to women, simply because they are women, full citizenship stature-equal opportunity to aspire, achieve, participate in and contribute to society based on their individual talents and capacities. . . .

 Without equating gender classifications, for all purposes, to classifications based on race or national origin,6 the Court, in post-Reed decisions, has carefully inspected official action that closes a door or denies opportunity to women (or to men). To summarize the Court's current directions for cases of official classification based on gender: Focusing on the differential treatment or denial of opportunity for which relief is sought, the reviewing court must determine whether the proffered justification is "exceedingly persuasive." The burden of justification is demanding and it rests entirely on the State. The State must show "at least that the [challenged] classification serves `important governmental objectives and that the discriminatory means employed' are `substantially related to the achievement of those objectives.'" The justification must be genuine, not hypothesized or invented post hoc in response to litigation. And it must not rely on overbroad generalizations about the different talents, capacities, or preferences of males and females.

 The heightened review standard our precedent establishes does not make sex a proscribed classification. Supposed "inherent differences" are no longer accepted as a ground for race or national origin classifications. Physical differences between men and women, however, are enduring: "[T]he two sexes are not fungible; a community made up exclusively of one [sex] is different from a community composed of both."

"Inherent differences" between men and women, we have come to appreciate, remain cause for celebration, but not for denigration of the members of either sex or for artificial constraints on an individual's opportunity. Sex classifications may be used to compensate women "for particular economic disabilities [they have] suffered," to "promot[e] equal employment opportunity," to advance full development of the talent and capacities of our Nation's people. But such classifications may not be used, as they once were, to create or perpetuate the legal, social, and economic inferiority of women.

 Neither recent nor distant history bears out Virginia's alleged pursuit of diversity through single-sex educational options. In 1839, when the State established VMI, a range of educational opportunities for men and women was scarcely contemplated. Higher education at the time was considered dangerous for women;9 reflecting widely held views about women's proper place, the Nation's first universities and colleges-for example, Harvard in Massachusetts, William and Mary in Virginia-admitted only men. VMI was not at all novel in this respect: In admitting no women, VMI followed the lead of the State's flagship school, the University of Virginia, founded in 1819.

. . .

    Virginia eventually provided for several women's seminaries and colleges. Farmville Female Seminary became a public institution in 1884. Two women's schools, Mary Washington College and James Madison University, were founded in 1908; another, Radford University, was founded in 1910. By the mid-1970's, all four schools had become coeducational.

 Debate concerning women's admission as undergraduates at the main university continued well past the century's midpoint. Familiar arguments were rehearsed. If women were admitted, it was feared, they "would encroach on the rights of men; there would be new problems of government, perhaps scandals; the old honor system would have to be changed; standards would be lowered to those of other coeducational schools; and the glorious reputation of the university, as a school for men, would be trailed in the dust."

 Ultimately, in 1970, "the most prestigious institution of higher education in Virginia," the University of Virginia, introduced coeducation and, in 1972, began to admit women on an equal basis with men. . . .

 Virginia describes the current absence of public single-sex higher education for women as "an historical anomaly." But the historical record indicates action more deliberate than anomalous: First, protection of women against higher education; next, schools for women far from equal in resources and stature to schools for men; finally, conversion of the separate schools to coeducation. The state legislature, prior to the advent of this controversy, had repealed "[a]ll Virginia statutes requiring individual institutions to admit only men or women." And in 1990, an official commission, "legislatively established to chart the future goals of higher education in Virginia," reaffirmed the policy "of affording broad access" while maintaining "autonomy and diversity." Significantly, the Commission reported:

"`Because colleges and universities provide opportunities for students to develop values and learn from role models, it is extremely important that they deal with faculty, staff, and students without regard to sex, race, or ethnic origin.'"

 Our 1982 decision in Mississippi Univ. for Women prompted VMI to reexamine its male-only admission policy. Virginia relies on that reexamination as a legitimate basis for maintaining VMI's single-sex character. A Mission Study Committee, appointed by the VMI Board of Visitors, studied the problem from October 1983 until May 1986, and in that month counseled against "change of VMI status as a single-sex college." Whatever internal purpose the Mission Study Committee served-and however well-meaning the framers of the report-we can hardly extract from that effort any state policy evenhandedly to advance diverse educational options. . . .

 In sum, we find no persuasive evidence in this record that VMI's male-only admission policy "is in furtherance of a state policy of `diversity.'" No such policy, the Fourth Circuit observed, can be discerned from the movement of all other public colleges and universities in Virginia away from single-sex education. That court also questioned "how one institution with autonomy, but with no authority over any other state institution, can give effect to a state policy of diversity among institutions." A purpose genuinely to advance an array of educational options, as the Court of Appeals recognized, is not served by VMI's historic and constant plan-a plan to "affor[d] a unique educational benefit only to males." However "liberally" this plan serves the State's sons, it makes no provision whatever for her daughters. That is not equal protection.

 The District Court forecast from expert witness testimony, and the Court of Appeals accepted, that coeducation would materially affect "at least these three aspects of VMI's program-physical training, the absence of privacy, and the adversative approach." And it is uncontested that women's admission would require accommodations, primarily in arranging housing assignments and physical training programs for female cadets. It is also undisputed, however, that "the VMI methodology could be used to educate women." The District Court even allowed that some women may prefer it to the methodology a women's college might pursue. "[S]ome women, at least, would want to attend [VMI] if they had the opportunity," the District Court recognized, and "some women," the expert testimony established, "are capable of all of the individual activities required of VMI cadets,". The parties, furthermore, agree that "some women can meet the physical standards [VMI] now impose[s] on men." In sum, as the Court of Appeals stated, "neither the goal of producing citizen soldiers," VMI's raison d'àtre, "nor VMI's implementing methodology is inherently unsuitable to women."

 In support of its initial judgment for Virginia, a judgment rejecting all equal protection objections presented by the United States, the District Court made "findings" on "gender-based developmental differences."  These "findings" restate the opinions of Virginia's expert witnesses, opinions about typically male or typically female "tendencies." Id., at 1434. For example, "[m]ales tend to need an atmosphere of adversativeness," while "[f]emales tend to thrive in a cooperative atmosphere." "I'm not saying that some women don't do well under [the] adversative model," VMI's expert on educational institutions testified, "undoubtedly there are some [women] who do"; but educational experiences must be designed "around the rule," this expert maintained, and not "around the exception."

 The United States does not challenge any expert witness estimation on average capacities or preferences of men and women. Instead, the United States emphasizes that time and again since this Court's turning point decision in Reed v. Reed, 404 U.S. 71 (1971), we have cautioned reviewing courts to take a "hard look" at generalizations or "tendencies" of the kind pressed by Virginia, and relied upon by the District Court. State actors controlling gates to opportunity, we have instructed, may not exclude qualified individuals based on "fixed notions concerning the roles and abilities of males and females." . . .

 It may be assumed, for purposes of this decision, that most women would not choose VMI's adversative method. As Fourth Circuit Judge Motz observed, however, in her dissent from the Court of Appeals' denial of rehearing en banc, it is also probable that "many men would not want to be educated in such an environment." (On that point, even our dissenting colleague might agree.) Education, to be sure, is not a "one size fits all" business. The issue, however, is not whether "women-or men-should be forced to attend VMI"; rather, the question is whether the State can constitutionally deny to women who have the will and capacity, the training and attendant opportunities that VMI uniquely affords.

 The notion that admission of women would downgrade VMI's stature, destroy the adversative system and, with it, even the school,11 is a judgment hardly proved,12 a prediction hardly different from other "self-fulfilling prophec[ies],"  once routinely used to deny rights or opportunities. When women first sought admission to the bar and access to legal education, concerns of the same order were expressed. For example, in 1876, the Court of Common Pleas of Hennepin County, Minnesota, explained why women were thought ineligible for the practice of law. Women train and educate the young, the court said, which

"forbids that they shall bestow that time (early and late) and labor, so essential in attaining to the eminence to which the true lawyer should ever aspire. It cannot therefore be said that the opposition of courts to the admission of females to practice . . . is to any extent the outgrowth of . . . `old fogyism[.]' . . . [I]t arises rather from a comprehension of the magnitude of the responsibilities connected with the successful practice of law, and a desire to grade up the profession."

"[t]he faculty . . . never maintained that women could not master legal learning . . . . No, its argument has been . . . more practical. If women were admitted to the Columbia Law School, [the faculty] said, then the choicer, more manly and red-blooded graduates of our great universities would go to the Harvard Law School!" The Nation, Feb. 18, 1925, p. 173.

 Women's successful entry into the federal military academies,13 and their participation in the Nation's military forces,14 indicate that Virginia's fears for the future of VMI may not be solidly grounded.15 The State's justification for excluding all women from "citizen-soldier" training for which some are qualified, in any event, cannot rank as "exceedingly persuasive," as we have explained and applied that standard.

 Virginia and VMI trained their argument on "means" rather than "end," and thus misperceived our precedent. Single-sex education at VMI serves an "important governmental objective," they maintained, and exclusion of women is not only "substantially related," it is essential to that objective. By this notably circular argument, the "straightforward" test Mississippi Univ. for Women described was bent and bowed.

 The State's misunderstanding and, in turn, the District Court's, is apparent from VMI's mission: to produce "citizen-soldiers," individuals

"`imbued with love of learning, confident in the functions and attitudes of leadership, possessing a high sense of public service, advocates of the American democracy and free enterprise system, and ready . . . to defend their country in time of national peril.'"

 Virginia chose not to eliminate, but to leave untouched, VMI's exclusionary policy. For women only, however, Virginia proposed a separate program, different in kind from VMI and unequal in tangible and intangible facilities.17 Having violated the Constitution's equal protection requirement, Virginia was obliged to show that its remedial proposal "directly address[ed] and relate[d] to" the violation, i.e., the equal protection denied to women ready, willing, and able to benefit from educational opportunities of the kind VMI offers. Virginia described VWIL as a "parallel program," and asserted that VWIL shares VMI's mission of producing "citizen-soldiers" and VMI's goals of providing "education, military training, mental and physical discipline, character . . . and leadership development." If the VWIL program could not "eliminate the discriminatory effects of the past," could it at least "bar like discrimination in the future"? See Louisiana, 380 U.S., at 154. A comparison of the programs said to be "parallel" informs our answer. In exposing the character of, and differences in, the VMI and VWIL programs, we recapitulate facts earlier presented.

 VWIL affords women no opportunity to experience the rigorous military training for which VMI is famed. . . . Instead, the VWIL program "deemphasize[s]" military education, and uses a "cooperative method" of education "which reinforces self-esteem,".

 VWIL students participate in ROTC and a "largely ceremonial" Virginia Corps of Cadets, but Virginia deliberately did not make VWIL a military institute. The VWIL House is not a military-style residence and VWIL students need not live together throughout the 4-year program, eat meals together, or wear uniforms during the school day. VWIL students thus do not experience the "barracks" life "crucial to the VMI experience," the spartan living arrangements designed to foster an "egalitarian ethic." "[T]he most important aspects of the VMI educational experience occur in the barracks," the District Court found, yet Virginia deemed that core experience nonessential, indeed inappropriate, for training its female citizen-soldiers.

 VWIL students receive their "leadership training" in seminars, externships, and speaker series, episodes and encounters lacking the "[p]hysical rigor, mental stress, . . . minute regulation of behavior, and indoctrination in desirable values" made hallmarks of VMI's citizen-soldier training. Kept away from the pressures, hazards, and psychological bonding characteristic of VMI's adversative training, VWIL students will not know the "feeling of tremendous accomplishment" commonly experienced by VMI's successful cadets.

 Virginia maintains that these methodological differences are "justified pedagogically," based on "important differences between men and women in learning and developmental needs," "psychological and sociological differences" Virginia describes as "real" and "not stereotypes." The Task Force charged with developing the leadership program for women, drawn from the staff and faculty at Mary Baldwin College, "determined that a military model and, especially VMI's adversative method, would be wholly inappropriate for educating and training most women." . . .

 As earlier stated, generalizations about "the way women are," estimates of what is appropriate for most women, no longer justify denying opportunity to women whose talent and capacity place them outside the average description. Notably, Virginia never asserted that VMI's method of education suits most men. It is also revealing that Virginia accounted for its failure to make the VWIL experience "the entirely militaristic experience of VMI" on the ground that VWIL "is planned for women who do not necessarily expect to pursue military careers." By that reasoning, VMI's "entirely militaristic" program would be inappropriate for men in general or as a group, for "[o]nly about 15% of VMI cadets enter career military service."

 In contrast to the generalizations about women on which Virginia rests, we note again these dispositive realties: VMI's "implementing methodology" is not "inherently unsuitable to women,"; "some women . . . do well under [the] adversative model,"; "some women, at least, would want to attend [VMI] if they had the opportunity,"; "some women are capable of all of the individual activities required of VMI cadets," and "can meet the physical standards [VMI] now impose[s] on men,". It is on behalf of these women that the United States has instituted this suit, and it is for them that a remedy must be crafted, a remedy that will end their exclusion from a state-supplied educational opportunity for which they are fit, a decree that will "bar like discrimination in the future."
 

 Mary Baldwin College, whose degree VWIL students will gain, enrolls first-year women with an average combined SAT score about 100 points lower than the average score for VMI freshmen. The Mary Baldwin faculty holds "significantly fewer Ph.D.'s," id., at 502, and receives substantially lower salaries, than the faculty at VMI.

 Mary Baldwin does not offer a VWIL student the range of curricular choices available to a VMI cadet. VMI awards baccalaureate degrees in liberal arts, biology, chemistry, civil engineering, electrical and computer engineering, and mechanical engineering. VWIL students attend a school that "does not have a math and science focus,"; they cannot take at Mary Baldwin any courses in engineering or the advanced math and physics courses VMI offers.

 For physical training, Mary Baldwin has "two multi-purpose fields" and "[o]ne gymnasium." VMI has "an NCAA competition level indoor track and field facility; a number of multi-purpose fields; baseball, soccer and lacrosse fields; an obstacle course; large boxing, wrestling and martial arts facilities; an 11-laps-to-the-mile indoor running course; an indoor pool; indoor and outdoor rifle ranges; and a football stadium that also contains a practice field and outdoor track."

 Although Virginia has represented that it will provide equal financial support for in-state VWIL students and VMI cadets, and the VMI Foundation has agreed to endow VWIL with $5.4625 million, the difference between the two schools' financial reserves is pronounced. Mary Baldwin's endowment, currently about $19 million, will gain an additional $35 million based on future commitments; VMI's current endowment, $131 million-the largest per-student endowment in the Nation-will gain $220 million.

 The VWIL student does not graduate with the advantage of a VMI degree. Her diploma does not unite her with the legions of VMI "graduates [who] have distinguished themselves" in military and civilian life. "[VMI] alumni are exceptionally close to the school," and that closeness accounts, in part, for VMI's success in attracting applicants. A VWIL graduate cannot assume that the "network of business owners, corporations, VMI graduates and non-graduate employers . . . interested in hiring VMI graduates," will be equally responsive to her search for employment . . .

 Virginia, in sum, while maintaining VMI for men only, has failed to provide any "comparable single-gender women's institution." Instead, the Commonwealth has created a VWIL program fairly appraised as a "pale shadow" of VMI in terms of the range of curricular choices and faculty stature, funding, prestige, alumni support and influence.

 Virginia's VWIL solution is reminiscent of the remedy Texas proposed 50 years ago, in response to a state trial court's 1946 ruling that, given the equal protection guarantee, African Americans could not be denied a legal education at a state facility. See Sweatt v. Painter, 339 U.S. 629 (1950). Reluctant to admit African Americans to its flagship University of Texas Law School, the State set up a separate school for Herman Sweatt and other black law students. As originally opened, the new school had no independent faculty or library, and it lacked accreditation. Nevertheless, the state trial and appellate courts were satisfied that the new school offered Sweatt opportunities for the study of law "substantially equivalent to those offered by the State to white students at the University of Texas."

 Before this Court considered the case, the new school had gained "a faculty of five full-time professors; a student body of 23; a library of some 16,500 volumes serviced by a full-time staff; a practice court and legal aid association; and one alumnus who ha[d] become a member of the Texas Bar." This Court contrasted resources at the new school with those at the school from which Sweatt had been excluded. The University of Texas Law School had a full-time faculty of 16, a student body of 850, a library containing over 65,000 volumes, scholarship funds, a law review, and moot court facilities.

 More important than the tangible features, the Court emphasized, are "those qualities which are incapable of objective measurement but which make for greatness" in a school, including "reputation of the faculty, experience of the administration, position and influence of the alumni, standing in the community, traditions and prestige."  Facing the marked differences reported in the Sweatt opinion, the Court unanimously ruled that Texas had not shown "substantial equality in the [separate] educational opportunities" the State offered.  Accordingly, the Court held, the Equal Protection Clause required Texas to admit African Americans to the University of Texas Law School. In line with Sweatt, we rule here that Virginia has not shown substantial equality in the separate educational opportunities the State supports at VWIL and VMI.
 

 The Fourth Circuit plainly erred in exposing Virginia's VWIL plan to a deferential analysis, for "all gender-based classifications today" warrant "heightened scrutiny." Valuable as VWIL may prove for students who seek the program offered, Virginia's remedy affords no cure at all for the opportunities and advantages withheld from women who want a VMI education and can make the grade. In sum, Virginia's remedy does not match the constitutional violation; the State has shown no "exceedingly persuasive justification" for withholding from women qualified for the experience premier training of the kind VMI affords.

 VMI, too, offers an educational opportunity no other Virginia institution provides, and the school's "prestige"-associated with its success in developing "citizen-soldiers"-is unequaled. Virginia has closed this facility to its daughters and, instead, has devised for them a "parallel program," with a faculty less impressively credentialed and less well paid, more limited course offerings, fewer opportunities for military training and for scientific specialization. VMI, beyond question, "possesses to a far greater degree" than the VWIL program "those qualities which are incapable of objective measurement but which make for greatness in a . . . school," including "position and influence of the alumni, standing in the community, traditions and prestige." Women seeking and fit for a VMI-quality education cannot be offered anything less, under the State's obligation to afford them genuinely equal protection.

 A prime part of the history of our Constitution, historian Richard Morris recounted, is the story of the extension of constitutional rights and protections to people once ignored or excluded.21 VMI's story continued as our comprehension of "We the People" expanded. There is no reason to believe that the admission of women capable of all the activities required of VMI cadets would destroy the Institute rather than enhance its capacity to serve the "more perfect Union."

 Chief Justice Rehnquist, concurring in judgment.

 The Court holds first that Virginia violates the Equal Protection Clause by maintaining the Virginia Military Institute's (VMI's) all-male admissions policy, and second that establishing the Virginia Women's Institute for Leadership (VWIL) program does not remedy that violation. While I agree with these conclusions, I disagree with the Court's analysis and so I write separately.

 While terms like "important governmental objective" and "substantially related" are hardly models of precision, they have more content and specificity than does the phrase "exceedingly persuasive justification." That phrase is best confined, as it was first used, as an observation on the difficulty of meeting the applicable test, not as a formulation of the test itself. . . . To avoid introducing potential confusion, I would have adhered more closely to our traditional, "firmly established" standard that a gender-based classification "must bear a close and substantial relationship to important governmental objectives."

 Our cases dealing with gender discrimination also require that the proffered purpose for the challenged law be the actual purpose. It is on this ground that the Court rejects the first of two justifications Virginia offers for VMI's single-sex admissions policy, namely, the goal of diversity among its public educational institutions. While I ultimately agree that the State has not carried the day with this justification, I disagree with the Court's method of analyzing the issue.

 VMI was founded in 1839, and, as the Court notes, admission was limited to men because under the then-prevailing view men, not women, were destined for higher education. However misguided this point of view may be by present-day standards, it surely was not unconstitutional in 1839. The adoption of the Fourteenth Amendment, with its Equal Protection Clause, was nearly 30 years in the future. The interpretation of the Equal Protection Clause to require heightened scrutiny for gender discrimination was yet another century away.

 Long after the adoption of the Fourteenth Amendment, and well into this century, legal distinctions between men and women were thought to raise no question under the Equal Protection Clause. The Court refers to our decision in Goesaert v. Cleary, 335 U.S. 466 (1948). Likewise representing that now abandoned view was Hoyt v. Florida, 368 U.S. 57 (1961), where the Court upheld a Florida system of jury selection in which men were automatically placed on jury lists, but women were placed there only if they expressed an affirmative desire to serve. The Court noted that despite advances in women's opportunities, the "woman is still regarded as the center of home and family life."

 Then, in 1971, we decided Reed v. Reed, 404 U.S. 71, which the Court correctly refers to as a seminal case. But its facts have nothing to do with admissions to any sort of educational institution. An Idaho statute governing the administration of estates and probate preferred men to women if the other statutory qualifications were equal. The statute's purpose, according to the Idaho Supreme Court, was to avoid hearings to determine who was better qualified as between a man and a woman both applying for letters of administration. This Court held that such a rule violated the Fourteenth Amendment because "a mandatory preference to members of either sex over members of the other, merely to accomplish the elimination of hearings" was an "arbitrary legislative choice forbidden by the Equal Protection Clause." Id., at 76. The brief opinion in Reed made no mention of either Goesaert or Hoyt.

 Even at the time of our decision in Reed v. Reed, therefore, Virginia and VMI were scarcely on notice that its holding would be extended across the constitutional board. They were entitled to believe that "one swallow doesn't make a summer" and await further developments. Those developments were 11 years in coming. In Mississippi Univ. for Women v. Hogan, 458 U.S. 718 (1982), a case actually involving a single-sex admissions policy in higher education, the Court held that the exclusion of men from a nursing program violated the Equal Protection Clause. This holding did place Virginia on notice that VMI's men-only admissions policy was open to serious question.

    . . .

 Before this Court, Virginia has sought to justify VMI's single-sex admissions policy primarily on the basis that diversity in education is desirable, and that while most of the public institutions of higher learning in the State are coeducational, there should also be room for single-sex institutions. I agree with the Court that there is scant evidence in the record that this was the real reason that Virginia decided to maintain VMI as men only.22 But, unlike the majority, I would consider only evidence that postdates our decision in Hogan, and would draw no negative inferences from the State's actions before that time. I think that after Hogan, the State was entitled to reconsider its policy with respect to VMI, and to not have earlier justifications, or lack thereof, held against it.

 Even if diversity in educational opportunity were the State's actual objective, the State's position would still be problematic. The difficulty with its position is that the diversity benefited only one sex; there was single-sex public education available for men at VMI, but no corresponding single-sex public education available for women. When Hogan placed Virginia on notice that VMI's admissions policy possibly was unconstitutional, VMI could have dealt with the problem by admitting women; but its governing body felt strongly that the admission of women would have seriously harmed the institution's educational approach. Was there something else the State could have done to avoid an equal protection violation? Since the State did nothing, we do not have to definitively answer that question.

 I do not think, however, that the State's options were as limited as the majority may imply. The Court cites, without expressly approving it, a statement from the opinion of the dissenting judge in the Court of Appeals, to the effect that the State could have "simultaneously opened single-gender undergraduate institutions having substantially comparable curricular and extra-curricular programs, funding, physical plant, administration and support services, and faculty and library resources." If this statement is thought to exclude other possibilities, it is too stringent a requirement. VMI had been in operation for over a century and a half, and had an established, successful and devoted group of alumni. No legislative wand could instantly call into existence a similar institution for women; and it would be a tremendous loss to scrap VMI's history and tradition. In the words of Grover Cleveland's second inaugural address, the State faced a condition, not a theory. And it was a condition that had been brought about, not through defiance of decisions construing gender bias under the Equal Protection Clause, but, until the decision in Hogan, a condition which had not appeared to offend the Constitution. Had Virginia made a genuine effort to devote comparable public resources to a facility for women, and followed through on such a plan, it might well have avoided an equal protection violation. I do not believe the State was faced with the stark choice of either admitting women to VMI, on the one hand, or abandoning VMI and starting from scratch for both men and women, on the other.

 But, as I have noted, neither the governing board of VMI nor the State took any action after 1982. If diversity in the form of single-sex, as well as coeducational, institutions of higher learning were to be available to Virginians, that diversity had to be available to women as well as to men.

 The dissent criticizes me for "disregarding the four all-women's private colleges in Virginia (generously assisted by public funds)." The private women's colleges are treated by the State exactly as all other private schools are treated, which includes the provision of tuition-assistance grants to Virginia residents. Virginia gives no special support to the women's single-sex education. But obviously, the same is not true for men's education. Had the State provided the kind of support for the private women's schools that it provides for VMI, this may have been a very different case. For in so doing, the State would have demonstrated that its interest in providing a single-sex education for men, was to some measure matched by an interest in providing the same opportunity for women.

 Virginia offers a second justification for the single-sex admissions policy: maintenance of the adversative method. I agree with the Court that this justification does not serve an important governmental objective. A State does not have substantial interest in the adversative methodology unless it is pedagogically beneficial. While considerable evidence shows that a single-sex education is pedagogically beneficial for some students, see 766 F. Supp., at 1414, and hence a State may have a valid interest in promoting that methodology, there is no similar evidence in the record that an adversative method is pedagogically beneficial or is any more likely to produce character traits than other methodologies.

 Accordingly, the remedy should not necessarily require either the admission of women to VMI, or the creation of a VMI clone for women. An adequate remedy in my opinion might be a demonstration by Virginia that its interest in educating men in a single-sex environment is matched by its interest in educating women in a single-sex institution. To demonstrate such, the State does not need to create two institutions with the same number of faculty PhD's, similar SAT scores, or comparable athletic fields. Nor would it necessarily require that the women's institution offer the same curriculum as the men's; one could be strong in computer science, the other could be strong in liberal arts. It would be a sufficient remedy, I think, if the two institutions offered the same quality of education and were of the same overall calibre.

 If a state decides to create single-sex programs, the state would, I expect, consider the public's interest and demand in designing curricula. And rightfully so. But the state should avoid assuming demand based on stereotypes; it must not assume a priori, without evidence, that there would be no interest in a women's school of civil engineering, or in a men's school of nursing.

 In the end, the women's institution Virginia proposes, VWIL, fails as a remedy, because it is distinctly inferior to the existing men's institution and will continue to be for the foreseeable future. VWIL simply is not, in any sense, the institution that VMI is. In particular, VWIL is a program appended to a private college, not a self-standing institution; and VWIL is substantially underfunded as compared to VMI. I therefore ultimately agree with the Court that Virginia has not provided an adequate remedy.

 Justice Scalia, dissenting.

 Today the Court shuts down an institution that has served the people of the Commonwealth of Virginia with pride and distinction for over a century and a half. To achieve that desired result, it rejects (contrary to our established practice) the factual findings of two courts below, sweeps aside the precedents of this Court, and ignores the history of our people. As to facts: it explicitly rejects the finding that there exist "gender-based developmental differences" supporting Virginia's restriction of the "adversative" method to only a men's institution, and the finding that the all-male composition of the Virginia Military Institute (VMI) is essential to that institution's character. As to precedent: it drastically revises our established standards for reviewing sex-based classifications. And as to history: it counts for nothing the long tradition, enduring down to the present, of men's military colleges supported by both States and the Federal Government.

 Much of the Court's opinion is devoted to deprecating the closed-mindedness of our forebears with regard to women's education, and even with regard to the treatment of women in areas that have nothing to do with education. Closed-minded they were-as every age is, including our own, with regard to matters it cannot guess, because it simply does not consider them debatable. The virtue of a democratic system with a First Amendment is that it readily enables the people, over time, to be persuaded that what they took for granted is not so, and to change their laws accordingly. That system is destroyed if the smug assurances of each age are removed from the democratic process and written into the Constitution. So to counterbalance the Court's criticism of our ancestors, let me say a word in their praise: they left us free to change. The same cannot be said of this most illiberal Court, which has embarked on a course of inscribing one after another of the current preferences of the society (and in some cases only the counter-majoritarian preferences of the society's law-trained elite) into our Basic Law. Today it enshrines the notion that no substantial educational value is to be served by an all-men's military academy-so that the decision by the people of Virginia to maintain such an institution denies equal protection to women who cannot attend that institution but can attend others. Since it is entirely clear that the Constitution of the United States-the old one-takes no sides in this educational debate, I dissent.

 I have no problem with a system of abstract tests such as rational-basis, intermediate, and strict scrutiny (though I think we can do better than applying strict scrutiny and intermediate scrutiny whenever we feel like it). Such formulas are essential to evaluating whether the new restrictions that a changing society constantly imposes upon private conduct comport with that "equal protection" our society has always accorded in the past. But in my view the function of this Court is to preserve our society's values regarding (among other things) equal protection, not to revise them; to prevent backsliding from the degree of restriction the Constitution imposed upon democratic government, not to prescribe, on our own authority, progressively higher degrees. For that reason it is my view that, whatever abstract tests we may choose to devise, they cannot supersede-and indeed ought to be crafted so as to reflect-those constant and unbroken national traditions that embody the people's understanding of ambiguous constitutional texts. More specifically, it is my view that "when a practice not expressly prohibited by the text of the Bill of Rights bears the endorsement of a long tradition of open, widespread, and unchallenged use that dates back to the beginning of the Republic, we have no proper basis for striking it down." The same applies, mutatis mutandis, to a practice asserted to be in violation of the post-Civil War Fourteenth Amendment. . . .

 The all-male constitution of VMI comes squarely within such a governing tradition. Founded by the Commonwealth of Virginia in 1839 and continuously maintained by it since, VMI has always admitted only men. And in that regard it has not been unusual. For almost all of VMI's more than a century and a half of existence, its single-sex status reflected the uniform practice for government-supported military colleges. Another famous Southern institution, The Citadel, has existed as a state-funded school of South Carolina since 1842. And all the federal military colleges-West Point, the Naval Academy at Annapolis, and even the Air Force Academy, which was not established until 1954-admitted only males for most of their history. Their admission of women in 1976 . . . came not by court decree, but because the people, through their elected representatives, decreed a change. In other words, the tradition of having government-funded military schools for men is as well rooted in the traditions of this country as the tradition of sending only men into military combat. The people may decide to change the one tradition, like the other, through democratic processes; but the assertion that either tradition has been unconstitutional through the centuries is not law, but politics-smuggled-into-law.

 And the same applies, more broadly, to single-sex education in general, which, as I shall discuss, is threatened by today's decision with the cut-off of all state and federal support. Government-run nonmilitary educational institutions for the two sexes have until very recently also been part of our national tradition. . . . These traditions may of course be changed by the democratic decisions of the people, as they largely have been.

 Today, however, change is forced upon Virginia, and reversion to single-sex education is prohibited nationwide, not by democratic processes but by order of this Court. Even while bemoaning the sorry, bygone days of "fixed notions" concerning women's education, the Court favors current notions so fixedly that it is willing to write them into the Constitution of the United States by application of custom-built "tests." This is not the interpretation of a Constitution, but the creation of one.

 Before I proceed to apply this standard to VMI, I must comment upon the manner in which the Court avoids doing so. Notwithstanding our above-described precedents and their "`firmly established principles,'" the United States urged us to hold in this case "that strict scrutiny is the correct constitutional standard for evaluating classifications that deny opportunities to individuals based on their sex." . . . The Court, while making no reference to the Government's argument, effectively accepts it.

 Although the Court in two places recites the test as stated in Hogan, which asks whether the State has demonstrated "that the classification serves important governmental objectives and that the discriminatory means employed are substantially related to the achievement of those objectives," the Court never answers the question presented in anything resembling that form. When it engages in analysis, the Court instead prefers the phrase "exceedingly persuasive justification" from Hogan. The Court's nine invocations of that phrase, and even its fanciful description of that imponderable as "the core instruction" of the Court's decisions in J. E. B. v. Alabama ex rel. T. B., 511 U.S. 127 (1994), and Hogan, supra, would be unobjectionable if the Court acknowledged that whether a "justification" is "exceedingly persuasive" must be assessed by asking "[whether] the classification serves important governmental objectives and [whether] the discriminatory means employed are substantially related to the achievement of those objectives." Instead, however, the Court proceeds to interpret "exceedingly persuasive justification" in a fashion that contradicts the reasoning of Hogan and our other precedents.

 That is essential to the Court's result, which can only be achieved by establishing that intermediate scrutiny is not survived if there are some women interested in attending VMI, capable of undertaking its activities, and able to meet its physical demands. . . .

 Only the amorphous "exceedingly persuasive justification" phrase, and not the standard elaboration of intermediate scrutiny, can be made to yield this conclusion that VMI's single-sex composition is unconstitutional because there exist several women (or, one would have to conclude under the Court's reasoning, a single woman) willing and able to undertake VMI's program. Intermediate scrutiny has never required a least-restrictive-means analysis, but only a "substantial relation" between the classification and the state interests that it serves. . . . There is simply no support in our cases for the notion that a sex-based classification is invalid unless it relates to characteristics that hold true in every instance.

 Not content to execute a de facto abandonment of the intermediate scrutiny that has been our standard for sex-based classifications for some two decades, the Court purports to reserve the question whether, even in principle, a higher standard (i.e., strict scrutiny) should apply. "The Court has," it says, "thus far reserved most stringent judicial scrutiny for classifications based on race or national origin . . ."; and it describes our earlier cases as having done no more than decline to "equat[e] gender classifications, for all purposes, to classifications based on race or national origin". The wonderful thing about these statements is that they are not actually false-just as it would not be actually false to say that "our cases have thus far reserved the `beyond a reasonable doubt' standard of proof for criminal cases," or that "we have not equated tort actions, for all purposes, to criminal prosecutions." But the statements are misleading, insofar as they suggest that we have not already categorically held strict scrutiny to be inapplicable to sex-based classifications. . . . And the statements are irresponsible, insofar as they are calculated to destabilize current law. Our task is to clarify the law-not to muddy the waters, and not to exact over-compliance by intimidation. The States and the Federal Government are entitled to know before they act the standard to which they will be held, rather than be compelled to guess about the outcome of Supreme Court peek-a-boo.

 The Court's intimations are particularly out of place because it is perfectly clear that, if the question of the applicable standard of review for sex-based classifications were to be regarded as an appropriate subject for reconsideration, the stronger argument would be not for elevating the standard to strict scrutiny, but for reducing it to rational-basis review. The latter certainly has a firmer foundation in our past jurisprudence: Whereas no majority of the Court has ever applied strict scrutiny in a case involving sex-based classifications, we routinely applied rational-basis review until the 1970's. And of course normal, rational-basis review of sex-based classifications would be much more in accord with the genesis of heightened standards of judicial review, the famous footnote in United States v. Carolene Products Co., 304 U.S. 144 (1938), which said (intimatingly) that we did not have to inquire in the case at hand

"whether prejudice against discrete and insular minorities may be a special condition, which tends seriously to curtail the operation of those political processes ordinarily to be relied upon to protect minorities, and which may call for a correspondingly more searching judicial inquiry."

 The evidence establishing that fact was overwhelming-indeed, "virtually uncontradicted" in the words of the court that received the evidence. As an initial matter, Virginia demonstrated at trial that "[a] substantial body of contemporary scholarship and research supports the proposition that, although males and females have significant areas of developmental overlap, they also have differing developmental needs that are deep-seated." While no one questioned that for many students a coeducational environment was nonetheless not inappropriate, that could not obscure the demonstrated benefits of single-sex colleges. . . .

    . . . "[I]n the light of this very substantial authority favoring single-sex education," the District Court concluded that "the VMI Board's decision to maintain an all-male institution is fully justified even without taking into consideration the other unique features of VMI's teaching and training." This finding alone, which even this Court cannot dispute, should be sufficient to demonstrate the constitutionality of VMI's all-male composition.

 But besides its single-sex constitution, VMI is different from other colleges in another way. It employs a "distinctive educational method," sometimes referred to as the "adversative, or doubting, model of education." "Physical rigor, mental stress, absolute equality of treatment, absence of privacy, minute regulation of behavior, and indoctrination in desirable values are the salient attributes of the VMI educational experience." No one contends that this method is appropriate for all individuals; education is not a "one size fits all" business. Just as a State may wish to support junior colleges, vocational institutes, or a law school that emphasizes case practice instead of classroom study, so too a State's decision to maintain within its system one school that provides the adversative method is "substantially related" to its goal of good education. Moreover, it was uncontested that "if the state were to establish a women's VMI-type [i.e., adversative] program, the program would attract an insufficient number of participants to make the program work,"; and it was found by the District Court that if Virginia were to include women in VMI, the school "would eventually find it necessary to drop the adversative system altogether,". Thus, Virginia's options were an adversative method that excludes women or no adversative method at all.

 There can be no serious dispute that, as the District Court found, single-sex education and a distinctive educational method "represent legitimate contributions to diversity in the Virginia higher education system." As a theoretical matter, Virginia's educational interest would have been best served (insofar as the two factors we have mentioned are concerned) by six different types of public colleges-an all-men's, an all-women's, and a coeducational college run in the "adversative method," and an all-men's, an all-women's, and a coeducational college run in the "traditional method." But as a practical matter, of course, Virginia's financial resources, like any State's, are not limitless, and the Commonwealth must select among the available options. Virginia thus has decided to fund, in addition to some 14 coeducational 4-year colleges, one college that is run as an all-male school on the adversative model: the Virginia Military Institute.

 Virginia did not make this determination regarding the make-up of its public college system on the unrealistic assumption that no other colleges exist. Substantial evidence in the District Court demonstrated that the Commonwealth has long proceeded on the principle that "`[h]igher education resources should be viewed as a whole-public and private'"-because such an approach enhances diversity and because "`it is academic and economic waste to permit unwarranted duplication.'" . . . It is thus significant that, whereas there are "four all-female private [colleges] in Virginia," there is only "one private all-male college," which "indicates that the private sector is providing for th[e] [former] form of education to a much greater extent that it provides for all-male education." In these circumstances, Virginia's election to fund one public all-male institution and one on the adversative model-and to concentrate its resources in a single entity that serves both these interests in diversity-is substantially related to the State's important educational interests.
 

 1.I have already pointed out the Court's most fundamental error, which is its reasoning that VMI's all-male composition is unconstitutional because "some women are capable of all of the individual activities required of VMI cadets," and would prefer military training on the adversative model. This unacknowledged adoption of what amounts to (at least) strict scrutiny is without antecedent in our sex-discrimination cases and by itself discredits the Court's decision.

 2.The Court suggests that Virginia's claimed purpose in maintaining VMI as an all-male institution-its asserted interest in promoting diversity of educational options-is not "genuin[e]," but is a pretext for discriminating against women. To support this charge, the Court would have to impute that base motive to VMI's Mission Study Committee, which conducted a 3-year study from 1983 to 1986 and recommended to VMI's Board of Visitors that the school remain all-male. The Committee, a majority of whose members consisted of non-VMI graduates, "read materials on education and on women in the military," "made site visits to single-sex and newly coeducational institutions" including West Point and the Naval Academy, and "considered the reasons that other institutions had changed from single-sex to coeducational status"; its work was praised as "thorough" in the accreditation review of VMI conducted by the Southern Association of Colleges and Schools. The Court states that "[w]hatever internal purpose the Mission Study Committee served-and however well-meaning the framers of the report-we can hardly extract from that effort any state policy evenhandedly to advance diverse educational options." But whether it is part of the evidence to prove that diversity was the Commonwealth's objective (its short report said nothing on that particular subject) is quite separate from whether it is part of the evidence to prove that anti-feminism was not. The relevance of the Mission Study Committee is that its very creation, its sober 3-year study, and the analysis it produced, utterly refute the claim that VMI has elected to maintain its all-male student-body composition for some misogynistic reason.

 The Court also supports its analysis of Virginia's "actual state purposes" in maintaining VMI's student body as all-male by stating that there is no explicit statement in the record "`in which the Commonwealth has expressed itself'" concerning those purposes. That is wrong on numerous grounds. First and foremost, in its implication that such an explicit statement of "actual purposes" is needed. The Court adopts, in effect, the argument of the United States that since the exclusion of women from VMI in 1839 was based on the "assumptions" of the time "that men alone were fit for military and leadership roles," and since "[b]efore this litigation was initiated, Virginia never sought to supply a valid, contemporary rationale for VMI's exclusionary policy," "[t]hat failure itself renders the VMI policy invalid." This is an unheard-of doctrine. Each state decision to adopt or maintain a governmental policy need not be accompanied-in anticipation of litigation and on pain of being found to lack a relevant state interest-by a lawyer's contemporaneous recitation of the State's purposes. The Constitution is not some giant Administrative Procedure Act, which imposes upon the States the obligation to set forth a "statement of basis and purpose" for their sovereign acts. The situation would be different if what the Court assumes to have been the 1839 policy had been enshrined and remained enshrined in legislation-a VMI charter, perhaps, pronouncing that the institution's purpose is to keep women in their place. But since the 1839 policy was no more explicitly recorded than the Court contends the present one is, the mere fact that today's Commonwealth continues to fund VMI "is enough to answer [the United States'] contention that the [classification] was the `accidental by-product of a traditional way of thinking about females.'"

    It is, moreover, not true that Virginia's contemporary reasons for maintaining VMI are not explicitly recorded. It is hard to imagine a more authoritative source on this subject than the 1990 Report of the Virginia Commission on the University of the 21st Century to the Governor and General Assembly (1990 Report). As the parties stipulated, that report "notes that the hallmarks of Virginia's educational policy are `diversity and autonomy.'" . . .

 The Court contends that "[a] purpose genuinely to advance an array of educational options . . . is not served" by VMI. It relies on the fact that all of Virginia's other public colleges have become coeducational. The apparent theory of this argument is that unless Virginia pursues a great deal of diversity, its pursuit of some diversity must be a sham. This fails to take account of the fact that Virginia's resources cannot support all possible permutations of schools, and of the fact that Virginia coordinates its public educational offerings with the offerings of in-state private educational institutions that the Commonwealth provides money for its residents to attend and otherwise assists-which include four women's colleges.

 Finally, the Court unreasonably suggests that there is some pretext in Virginia's reliance upon decentralized decisionmaking to achieve diversity-its granting of substantial autonomy to each institution with regard to student-body composition and other matters. The Court adopts the suggestion of the Court of Appeals that it is not possible for "one institution with autonomy, but with no authority over any other state institution, [to] give effect to a state policy of diversity among institutions." If it were impossible for individual human beings (or groups of human beings) to act autonomously in effective pursuit of a common goal, the game of soccer would not exist. And where the goal is diversity in a free market for services, that tends to be achieved even by autonomous actors who act out of entirely selfish interests and make no effort to cooperate. Each Virginia institution, that is to say, has a natural incentive to make itself distinctive in order to attract a particular segment of student applicants. And of course none of the institutions is entirely autonomous; if and when the legislature decides that a particular school is not well serving the interest of diversity-if it decides, for example, that a men's school is not much needed-funding will cease.

 3.In addition to disparaging Virginia's claim that VMI's single-sex status serves a state interest in diversity, the Court finds fault with Virginia's failure to offer education based on the adversative training method to women. It dismisses the District Court's "`findings' on `gender-based developmental differences'" on the ground that "[t]hese `findings' restate the opinions of Virginia's expert witnesses, opinions about typically male or typically female `tendencies.'" . . .

 Ultimately, in fact, the Court does not deny the evidence supporting these findings. It instead makes evident that the parties to this case could have saved themselves a great deal of time, trouble, and expense by omitting a trial. The Court simply dispenses with the evidence submitted at trial-it never says that a single finding of the District Court is clearly erroneous-in favor of the Justices' own view of the world, which the Court proceeds to support with (1) references to observations of someone who is not a witness, nor even an educational expert, nor even a judge who reviewed the record or participated in the judgment below, but rather a judge who merely dissented from the Court of Appeals' decision not to rehear this case en banc (2) citations of nonevidentiary materials such as amicus curiae briefs filed in this Court, and (3) various historical anecdotes designed to demonstrate that Virginia's support for VMI as currently constituted reminds the Justices of the "bad old days".

 It is not too much to say that this approach to the case has rendered the trial a sham. But treating the evidence as irrelevant is absolutely necessary for the Court to reach its conclusion. . . .

 4.The Court contends that Virginia, and the District Court, erred, and "misperceived our precedent," by "train[ing] their argument on `means' rather than `end,'". The Court focuses on "VMI's mission," which is to produce individuals "imbued with love of learning, confident in the functions and attitudes of leadership, possessing a high sense of public service, advocates of the American democracy and free enterprise system, and ready . . . to defend their country in time of national peril." "Surely," the Court says, "that goal is great enough to accommodate women."

 This is law-making by indirection. What the Court describes as "VMI's mission" is no less the mission of all Virginia colleges. Which of them would the Old Dominion continue to fund if they did not aim to create individuals "imbued with love of learning, etc." right down to being ready "to defend their country in time of national peril"? It can be summed up as "learning, leadership, and patriotism." To be sure, those general educational values are described in a particularly martial fashion in VMI's mission statement, in accordance with the military, adversative, and all-male character of the institution. But imparting those values in that fashion-i.e., in a military, adversative, all-male environment-is the distinctive mission of VMI. And as I have discussed . . . , that mission is not "great enough to accommodate women." The Court's analysis at least has the benefit of producing foreseeable results. Applied generally, it means that whenever a State's ultimate objective is "great enough to accommodate women" (as it always will be), then the State will be held to have violated the Equal Protection Clause if it restricts to men even one means by which it pursues that objective-no matter how few women are interested in pursuing the objective by that means, no matter how much the single-sex program will have to be changed if both sexes are admitted, and no matter how beneficial that program has theretofore been to its participants.

 5.The Court argues that VMI would not have to change very much if it were to admit women. The principal response to that argument is that it is irrelevant: If VMI's single-sex status is substantially related to the government's important educational objectives, as I have demonstrated above and as the Court refuses to discuss, that concludes the inquiry. There should be no debate in the federal judiciary over "how much" VMI would be required to change if it admitted women and whether that would constitute "too much" change.

 But if such a debate were relevant, the Court would certainly be on the losing side. . . . Changes that the District Court's detailed analysis found would be required include new allowances for personal privacy in the barracks, such as locked doors and coverings on windows, which would detract from VMI's approach of regulating minute details of student behavior, "contradict the principle that everyone is constantly subject to scrutiny by everyone else," and impair VMI's "total egalitarian approach" under which every student must be "treated alike"; changes in the physical training program, which would reduce "[t]he intensity and aggressiveness of the current program"; and various modifications in other respects of the adversative training program which permeates student life. As the Court of Appeals summarized it, "the record supports the district court's findings that at least these three aspects of VMI's program-physical training, the absence of privacy, and the adversative approach-would be materially affected by coeducation, leading to a substantial change in the egalitarian ethos that is a critical aspect of VMI's training."

 In the face of these findings by two courts below, amply supported by the evidence, and resulting in the conclusion that VMI would be fundamentally altered if it admitted women, this Court simply pronounces that "[t]he notion that admission of women would downgrade VMI's stature, destroy the adversative system and, with it, even the school, is a judgment hardly proved." . . .

 6.Finally, the absence of a precise "all-women's analogue" to VMI is irrelevant. . . .

 Although there is no precise female-only analogue to VMI, Virginia has created during this litigation the Virginia Women's Institute for Leadership (VWIL), a state-funded all-women's program run by Mary Baldwin College. I have thus far said nothing about VWIL because it is, under our established test, irrelevant, so long as VMI's all-male character is "substantially related" to an important state goal. But VWIL now exists, and the Court's treatment of it shows how far-reaching today's decision is.

 VWIL was carefully designed by professional educators who have long experience in educating young women. The program rejects the proposition that there is a "difference in the respective spheres and destinies of man and woman," and is designed to "provide an all-female program that will achieve substantially similar outcomes [to VMI's] in an all-female environment." . . .

    It is worth noting that none of the United States' own experts in the remedial phase of this case was willing to testify that VMI's adversative method was an appropriate methodology for educating women. This Court, however, does not care. Even though VWIL was carefully designed by professional educators who have tremendous experience in the area, and survived the test of adversarial litigation, the Court simply declares, with no basis in the evidence, that these professionals acted on "`overbroad' generalizations."

. . .
 

 The Supreme Court of the United States does not sit to announce "unique" dispositions. Its principal function is to establish precedent-that is, to set forth principles of law that every court in America must follow. As we said only this Term, we expect both ourselves and lower courts to adhere to the "rationale upon which the Court based the results of its earlier decisions." That is the principal reason we publish our opinions.

 And the rationale of today's decision is sweeping: for sex-based classifications, a redefinition of intermediate scrutiny that makes it indistinguishable from strict scrutiny. Indeed, the Court indicates that if any program restricted to one sex is "uniqu[e]," it must be opened to members of the opposite sex "who have the will and capacity" to participate in it. . . .

 In any event, regardless of whether the Court's rationale leaves some small amount of room for lawyers to argue, it ensures that single-sex public education is functionally dead. The costs of litigating the constitutionality of a single-sex education program, and the risks of ultimately losing that litigation, are simply too high to be embraced by public officials. . . . No state official in his right mind will buy such a high-cost, high-risk lawsuit by commencing a single-sex program. The enemies of single-sex education have won; by persuading only seven Justices (five would have been enough) that their view of the world is enshrined in the Constitution, they have effectively imposed that view on all 50 States.

 This is especially regrettable because, as the District Court here determined, educational experts in recent years have increasingly come to "suppor[t] [the] view that substantial educational benefits flow from a single-gender environment, be it male or female, that cannot be replicated in a coeducational setting." "The evidence in th[is] case," for example, "is virtually uncontradicted" to that effect. Until quite recently, some public officials have attempted to institute new single-sex programs, at least as experiments. In 1991, for example, the Detroit Board of Education announced a program to establish three boys-only schools for inner-city youth; it was met with a lawsuit, a preliminary injunction was swiftly entered by a District Court that purported to rely on Hogan, and the Detroit Board of Education voted to abandon the litigation and thus abandon the plan. Today's opinion assures that no such experiment will be tried again.
 

 The Court adverts to private single-sex education only briefly, and only to make the assertion (mentioned above) that "[w]e address specifically and only an educational opportunity recognized by the District Court and the Court of Appeals as `unique.'" As I have already remarked, that assurance assures nothing, unless it is to be taken as a promise that in the future the Court will disclaim the reasoning it has used today to destroy VMI. The Government, in its briefs to this Court, at least purports to address the consequences of its attack on VMI for public support of private single-sex education. It contends that private colleges which are the direct or indirect beneficiaries of government funding are not thereby necessarily converted into state actors to which the Equal Protection Clause is then applicable. That is true. It is also virtually meaningless.

 The issue will be not whether government assistance turns private colleges into state actors, but whether the government itself would be violating the Constitution by providing state support to single-sex colleges. For example, in Norwood v. Harrison, 413 U.S. 455 (1973), we saw no room to distinguish between state operation of racially segregated schools and state support of privately run segregated schools. . . . When the Government was pressed at oral argument concerning the implications of these cases for private single-sex education if government-provided single-sex education is unconstitutional, it stated that the implications will not be so disastrous, since States can provide funding to racially segregated private schools, "depend[ing] on the circumstances,". I cannot imagine what those "circumstances" might be, and it would be as foolish for private-school administrators to think that that assurance from the Justice Department will outlive the day it was made, as it was for VMI to think that the Justice Department's "unequivoca[l]" support for an intermediate-scrutiny standard in this case would survive the Government's loss in the courts below.

 The only hope for state-assisted single-sex private schools is that the Court will not apply in the future the principles of law it has applied today. That is a substantial hope, I am happy and ashamed to say. After all, did not the Court today abandon the principles of law it has applied in our earlier sex-classification cases? And does not the Court positively invite private colleges to rely upon our ad-hocery by assuring them this case is "unique"? I would not advise the foundation of any new single-sex college (especially an all-male one) with the expectation of being allowed to receive any government support; but it is too soon to abandon in despair those single-sex colleges already in existence. It will certainly be possible for this Court to write a future opinion that ignores the broad principles of law set forth today, and that characterizes as utterly dispositive the opinion's perceptions that VMI was a uniquely prestigious all-male institution, conceived in chauvinism, etc., etc. I will not join that opinion.

 Justice Brandeis said it is "one of the happy incidents of the federal system that a single courageous State may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country." But it is one of the unhappy incidents of the federal system that a self-righteous Supreme Court, acting on its Members' personal view of what would make a "more perfect Union," (a criterion only slightly more restrictive than a "more perfect world"), can impose its own favored social and economic dispositions nationwide. As today's disposition, and others this single Term, show, this places it beyond the power of a "single courageous State," not only to introduce novel dispositions that the Court frowns upon, but to reintroduce, or indeed even adhere to, disfavored dispositions that are centuries old. See, e.g., BMW of North America, Inc. v. Gore, 517 U. S. ___ (1996); Romer v. Evans, 517 U. S. ___ (1996). The sphere of self-government reserved to the people of the Republic is progressively narrowed.

 In the course of this dissent, I have referred approvingly to the opinion of my former colleague, Justice Powell, in Mississippi Univ. for Women v. Hogan, 458 U.S. 718 (1982). Many of the points made in his dissent apply with equal force here-in particular, the criticism of judicial opinions that purport to be "narro[w]" but whose "logic" is "sweepin[g]." But there is one statement with which I cannot agree. Justice Powell observed that the Court's decision in Hogan, which struck down a single-sex program offered by the Mississippi University for Women, had thereby "[l]eft without honor . . . an element of diversity that has characterized much of American education and enriched much of American life." Id., at 735. Today's decision does not leave VMI without honor; no court opinion can do that.

 In an odd sort of way, it is precisely VMI's attachment to such old-fashioned concepts as manly "honor" that has made it, and the system it represents, the target of those who today succeed in abolishing public single-sex education. The record contains a booklet that all first-year VMI students (the so-called "rats") were required to keep in their possession at all times. Near the end there appears the following period-piece, entitled "The Code of a Gentleman":

"Without a strict observance of the fundamental Code of Honor, no man, no matter how `polished,' can be considered a gentleman. The honor of a gentleman demands the inviolability of his word, and the incorruptibility of his principles. He is the descendant of the knight, the crusader; he is the defender of the defenseless and the champion of justice . . . or he is not a Gentleman.

 Does not discuss his family affairs in public or with acquaintances.
 Does not speak more than casually about his girl friend.
 Does not go to a lady's house if he is affected by alcohol. He is temperate in the use of alcohol.
 Does not lose his temper; nor exhibit anger, fear, hate, embarrassment, ardor or hilarity in public.
 Does not hail a lady from a club window.
 A gentleman never discusses the merits or demerits of a lady.
 Does not mention names exactly as he avoids the mention of what things cost.
 Does not borrow money from a friend, except in dire need. Money borrowed is a debt of honor, and must be repaid as promptly as possible. Debts incurred by a deceased parent, brother, sister or grown child are assumed by honorable men as a debt of honor.
 Does not display his wealth, money or possessions.
 Does not put his manners on and off, whether in the club or in a ballroom. He treats people with courtesy, no matter what their social position may be.
 Does not slap strangers on the back nor so much as lay a finger on a lady.
 Does not `lick the boots of those above' nor `kick the face of those below him on the social ladder.'
 Does not take advantage of another's helplessness or ignorance and assumes that no gentleman will take advantage of him.
 A Gentleman respects the reserves of others, but demands that others respect those which are his.
 A Gentleman can become what he wills to be. . ."

 I do not know whether the men of VMI lived by this Code; perhaps not. But it is powerfully impressive that a public institution of higher education still in existence sought to have them do so. I do not think any of us, women included, will be better off for its destruction.