Affirmed and remanded by published opinion. Judge NIEMEYER wrote the opinion, in which Senior Judge WARD joined. Senior Judge PHILLIPS wrote a separate dissenting opinion.
OPINION
NIEMEYER, Circuit Judge:At issue is the important question of whether a state may sponsor single-gender education without violating the Equal Protection Clause of the Fourteenth Amendment.
In United States v. Commonwealth of Virginia, (VMI I), 976 F.2d 890 (4th Cir.1992), we concluded that single-gender education was “pedagogically justifiable,” id. at 897, and the United States has acknowledged in this case that state sponsorship of single-gender education, if provided to both genders, is not per se a denial of equal protection. Even though single-gender college education yields benefits to both genders, it nevertheless has the secondary effect of excluding men from the women’s college and women from the men’s college, an effect that becomes yet more complicated when the programs at the two colleges differ to some degree.
We must decide now whether the Commonwealth of Virginia’s proposal (1) to continue to provide a single-gender military-type college education for men at the Virginia Military Institute (VMI), (2) to provide, beginning in 1995, a single-gender education with special leadership training for women at Mary Baldwin College, and (3) to continue to provide other forms of college education, including military training, for both men and women at other colleges and universities in the state is constitutionally permissible. After applying a heightened intermediate scrutiny test specially tailored to the circumstances before us and imposing specific performance criteria on the implementation of Virginia’s proposal, we affirm the district court’s judgment approving the proposal.
I
VMI, established by the Commonwealth of Virginia in 1839 as a four-year military college, has a current enrollment of approximately 1,300 men. The college has always admitted only males and, through an adver-sative military-type training, it seeks to graduate them as “ ‘citizen-soldiers, educated and honorable men who are suited for leadership in civilian life and who can provide military leadership when necessary.’ ” VMI I, 976 F.2d at 893. In VMI I, we affirmed the district court’s factual findings, based on studies in evidence, that such a single-gender education is pedagogically justifiable, both for males and females. We concluded:
*1233It is not the maleness, as distinguished from femaleness, that provides justification for the program. It is the homogeneity of gender in the process, regardless of which sex is considered, that has been shown to be related to the essence of the education and training at VMI.
Id. at 897.
We also affirmed findings of fact that coeducation would destroy aspects of VMI’s program which he near the core of its holistic system and that the admission of women therefore would deny them the very benefit they sought by their admission. The district court found that coeducation would require fundamental changes (1) to the adversative method which pits male against male because that method would not produce the same results when a male is set against a female; (2) to the absence of privacy which was found to be essential to the leveling process; and (3) to physical training, requiring VMI to adopt, as was required at the U.S. military academies, a dual-track program for men and women in order to achieve equality in effect. We concluded that coeducation at VMI would thus
deny those women the very opportunity they sought because the unique characteristics of VMI’s program would be destroyed by coeducation. The Catch-22 is that women are denied the opportunity when excluded from VMI and cannot be given the opportunity by admitting them, because the change caused by their admission would destroy the opportunity.
Id. at 897 (footnote omitted).
In view of these findings, we did not direct the Commonwealth of Virginia to change VMI to a coeducational college, but we did find that its failure to offer women comparable benefits constituted a violation of the Equal Protection Clause of the Fourteenth Amendment. We remanded the case to the district court, directing it to require Virginia and the other defendants to formulate, adopt, and oversee the implementation of a remedial plan. In giving Virginia the opportunity to select its course to correct the Fourteenth Amendment violations, we did not suggest any particular remedy, but allowed that Virginia might properly decide to alter the program and admit women to VMI, or establish parallel institutions or parallel programs, or abandon state support, leaving VMI the option to pursue its own policies as a private institution.
On remand, Virginia designed a proposal to implement a parallel program at Mary Baldwin College providing women with single-gender education, coupled with special leadership training. Following a trial on the appropriateness of the remedy, the district court approved the plan and directed Virginia “to proceed with all deliberate speed in implementing the Plan and to have the Plan operational for the academic year commencing in the Fall of 1995.” United States v. Commonwealth of Virginia, 852 F.Supp. 471, 485 (W-D-Va.1994). The court retained jurisdiction to supervise implementation of the plan and required a status report every six months.
The plan approved by the district court provides for Virginia to establish with state funds the Virginia Women’s Institute for Leadership (VWIL) as part of the undergraduate program at the otherwise privately funded Mary Baldwin College, a women’s liberal arts college founded in 1842 in Staun-ton, Virginia, about 35 miles from VMI. The plan is the product of a task force, chaired by Dr. James D. Lott, Dean of Mary Baldwin College, which set as its goal the task of designing a program at Mary Baldwin College to produce “citizen-soldiers who are educated and honorable women, prepared for varied work of civil life, qualified to serve in the armed forces, imbued with love of learning, confident in the functions and attitudes of leadership, and possessing a high sense of public service.” Because its mission is similar to VMI’s mission, VWIL would have its students pursue the same five goals as those pursued at VMI: education, military training, mental and physical discipline, character development, and leadership development. In designing the program at Mary Baldwin College, however, the task force concluded that aspects of VMI’s military model, especially the adversative method, would not be effective for women as a group, even though the task force concluded that some women would be suited to and interested in experi-*1234ending a “women’s VMI.” The task force concluded instead that its mission and goals could better be achieved by designing a program which deemphasized the military methods associated with the “rat line,” see VMI I, 976 F.2d at 893, utilizing instead a structured environment emphasizing leadership training.
In addition to the standard bachelor of arts program offered at Mary Baldwin College, VWIL students would be required to complete, as a “minor,” core and elective courses in leadership. A student in the VWIL program would be required to take courses in leadership communications; theories of leadership; ethics, community, and leadership; and a leadership seminar or semester of independent research on a topic relevant to women and leadership. Students would also be required to participate in Saturday seminars sponsored by upperclass students on designated subjects. Outside of the classroom, students would be required to complete a leadership externship during which they would work off campus in the public or private sector for up to one semester and to participate in a speaker series in which each VWIL class would be responsible for bringing outstanding leaders to speak on campus. Finally, all VWIL students would be required to organize and carry out community service projects.
While students at VWIL would be required to participate in four years of ROTC and in an ROTC summer camp, VWIL would not be organized under the pervasive military regimen that exists at VMI. Nevertheless, in addition to standard ROTC training, the students would conduct “leadership laboratory activities” which might incorporate aspects of military training, and they would participate in a newly-established Virginia Corps of Cadets, a uniformed military corps comprised of the all-female VWIL, the all-male VMI, and the coeducational Virginia Tech ROTC corps. The Virginia Corps of Cadets would be largely ceremonial.
Finally, VWIL students would be required to take and pass eight semesters of physical education, a portion of which would be devoted to health education courses. These programs would include athletics, physical training and a “cooperative confidence budding program” to be held twice a week.
The VWIL program would be implemented at Mary Baldwin College with its faculty, although VMI faculty would conduct some ROTC training and teach some ROTC courses at Mary Baldwin College. The program would be funded by the Commonwealth of Virginia, providing a per student payment equal to the current annual appropriation paid per cadet at VMI. The program, which task force members expect would have about 25 to 30 students in the first year, would also be given a permanent endowment of $5.46 million. The out-of-pocket expenses for students to attend VWIL is expected to be no greater than those of students attending VMI, and VWIL students would be eligible for the same financial aid programs as are available to VMI cadets.
The experts for both sides acknowledge that the proposed VWIL program differs from VMI in methodology since VWIL would not rely on the pervasive military life and adversative methods to achieve its goals. Members of the task force, who are professionals in education, testified that the different approach was selected principally to address the different educational needs of most women. Dr. Heather Anne Wilson, a member of the task force, summarized the thinking, stating that “the VMI model is based on the premise that young men come with [an] inflated sense of self-efficacy that must [be] knocked down and rebuilt_ What [women] need is a system that builds their sense of self-efficacy through meeting challenges, developing self-discipline, meeting rigor and dealing with it, and having successes.” Mary Baldwin College, which participated actively in the design of the VWIL program, observed in its amicus brief regarding the differences between the methodology used at VMI and that proposed for VWIL:
It would have been possible to design the VWIL program to more closely resemble VMI, with identical physical fitness standards and adversative techniques associated with the rat line. Such a program would have been easier to design and to defend against the arguments raised by the government and its amici. But it *1235would have been a paper program, with no real prospect of successful implementation. [Mary Baldwin College] believes it would be professionally irresponsible to compromise student welfare by designing a program to meet litigation objectives instead of student needs.
While the task force did not conduct any scientific survey on demand for the proposed VWIL program, or alternatively for a women’s YMI, several members expressed the opinion, based on some field data, that demand would be “significant” for VWIL but not for a women’s VMI, and some expressed doubt that enough women would be interested in a women’s VMI to make it work.
II
The United States contends that the remedial program offered by the Commonwealth of Virginia does not meet the requirements of the Fourteenth Amendment’s Equal Protection Clause. It states that “[the proposed remedy] does not correct the constitutional violation, i.e., the denial to women of VMI’s unique educational methodology. As the district court recognized, the program ‘differs substantially’ from the educational program offered at VMI.” At oral argument the United States argued that any parallel program would have to be “identical” in substance and methodology to that of VMI. The United States maintains further that by not offering coeducation at VMI, the Commonwealth of Virginia is relying on false stereotypes and generalizations “that women are not tough enough to succeed in VMI’s rigorous, military-style program.” As the United States summarized its position:
[T]he fact remains that men have [this] special educational opportunity available to them and women do not, and that as a result VMI graduates have been very successful in both public and private careers. This suit was brought on behalf of those women who want to go to VMI precisely because it is such a demanding and challenging school. The remedial plan approved by the district court does nothing for them.
The United States urges that we enter an order directing Virginia to admit women to VMI as the only remedy for correcting the past constitutional violation.
Ill
Equal protection of the law requires that persons similarly circumstanced be treated alike, Reed v. Reed, 404 U.S. 71, 76 (1971), but equal protection does not deny states the power “to treat different classes of persons in different ways.” Id. at 75. When the state classifies by defining a group to whom a regulation applies or a benefit is conferred, the classification “‘must be reasonable, not arbitrary, and must rest upon some ground of difference having a fair and substantial relation to the object of the legislation _’” Id. at 76 (quoting Royster Guano Co. v. Virginia, 253 U.S. 412, 415 (1920)). In striking down a state statute that preferred males over females as administrators of wills in order to further governmental efficiency, the Court provided the seed for the formulation of a test utilizing an intermediate level of scrutiny for state regulations that classify by gender. While Reed implicitly applied a heightened level of scrutiny, the formulation of this standard came later. As this test has finally been articulated, to withstand this level of scrutiny, “classifications by gender must serve important governmental objectives and must be substantially related to achievement of those objectives.” Orr v. Orr, 440 U.S. 268, 279 (1979) (internal quotations omitted); see also Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982).
In Hogan the Court outlined a two-step process that inquires (1) whether the state’s objective is “legitimate and important,” and (2) whether “the requisite direct, substantial relationship between objective and means is present.” Id. at 725. The Court explained that tailoring the means to fit the legitimate and important purpose is necessary to assure that the classification is the product of “reasoned analysis” rather than the “mechanical application of traditional, often inaccurate, assumptions about the proper roles of men and women.” Id. at 725-26. In Hogan, the Court held unconstitutional Mississippi’s women-only admissions policy of a state supported nursing school. In doing so, the *1236Court relied on the first prong of its articulated test, finding that the state’s purported objective to “compensate[ ] for [past] discrimination against women,” id. at 727, was in effect an effort to “ ‘protect’ members of one gender because they are presumed to suffer from an inherent handicap or to be innately inferior.” Id. at 725. The state’s purported justification, the Court observed, perpetuated an archaic and stereotyped view of women as nurses. Thus, affirmative action in favor of women based on an outdated and unsupported assertion that such protection was needed was held not to be an important governmental objective. Cf. Kirchberg v. Feenstra, 450 U.S. 455 (1981) (holding that a state statute, which preserves the husband as “the head and master of the [marriage] partnership or community of gains” by providing the husband, and not the wife, with the right to unilaterally dispose of jointly held property, was not a legitimate and important state objective). The Court in Hogan did not decide, however, and indeed appears deliberately to have left open, the question of whether states could provide single-gender education in other circumstances. See 458 U.S. at 720 n. 1.
In undertaking the first step of the Hogan analysis to determine whether the state’s objective is “legitimate and important,” a court should not substitute its priorities of value over those established by the democratically chosen branch. To remain true to its constitutional role and avoid the pitfalls of a substantive evaluation of proper governmental objectives, which would amount to a “substantive equal protection analysis,” a court should, at this step, deferentially consider only whether the regulation is important to a legitimate governmental purpose. The cautious approach to this first prong of intermediate scrutiny effectively redirects the court’s focus on evaluating the state’s means for obtaining its objective, which is the second step to the Hogan analysis. Giving greater scrutiny to the selection of means than to the proffered objective recognizes an appropriate deference to legislative will and at the same time assures that the legislature does not accomplish its objectives through an unequal application of the law. The substantive equal protection portion of the analysis cannot, however, be entirely diminished, for the courts can never approve a pernicious legislative purpose or one that does not comport with traditional notions of the proper role of government. The proper relative balance under this “procedural equal protection analysis” thus results in a court scrutinizing closely the procedural mechanism adopted by the legislature to accomplish its purpose and determining whether the means selected fits that purpose and bears a direct and substantial relationship to it. Cf. Faulkner v. Jones, 10 F.3d 226, 230 (4th Cir.1993) (“A regulatory classification which is made for a purpose unrelated to the purpose of the regulation, or which is broader than that appropriate for the regulation, may reveal prejudice and define discrimination.”).
There is ample support for directing the court’s attention from a substantive equal protection analysis to a more procedural analysis. In Craig v. Boren, 429 U.S. 190, 199-200 (1976), the Court recognized, without a significant substantive evaluation, that the protection of public health and safety, which formed the basis for Oklahoma’s statutes prohibiting the sale of “3.2% beer” to males under 21 and females under 18, “represents an important function of state and local governments.” But the Court found the statute unconstitutional nevertheless because the classification by gender was not shown to serve a sufficient role in achieving the state’s objective. Similarly in Reed, the Court agreed, again without any significant substantive evaluation, that establishing an efficient probate process was a legitimate governmental objective. But it nevertheless held that selecting males over females as administrators of wills was not a means that bore a direct and substantial relationship to the state’s objective. See 404 U.S. at 76-77. See also Orr v. Orr, 440 U.S. 268, 280 (1979) (readily acknowledging as legitimate the state’s purpose of providing assistance to needy spouses but finding unconstitutional the means that required husbands, but not wives, to pay alimony upon divorce).
Accordingly, under the intermediate level of scrutiny of a statute or program that *1237classifies by gender, the analysis begins with, the limited inquiry into whether the state objective is both consistent with a legitimate governmental role and important in serving that role. Thereafter it must shift to an inquiry of heightened scrutiny into whether the classification “substantially and directly furthers” that objective.
Application of this traditional test, however, to a ease where the classification is not directed per se at men or women, but at homogeneity of gender, presents a unique problem, because once the state’s objective is found to be an important one, the classification by gender is by definition necessary for accomplishing the objective and might thereby bypass any equal protection scrutiny. The second prong of the test thus would provide little or no scrutiny of the effect of a classification directed at homogeneity of gender. Thus, in order to measure the legitimacy of a classification based on homogeneity of gender against the Equal Protection Clause, we conclude that we must take the additional step of carefully weighing the alternatives available to members of each gender denied benefits by the classification.
To achieve the equality of treatment demanded by the Equal Protection Clause, the alternatives left available to each gender by a classification based on a homogeneity of gender need not be the same, but they must be substantively comparable so that, in the end, we cannot conclude that the value of the benefits provided by the state to one gender tends, by comparison to the benefits provided to the other, to lessen the dignity, respect, or societal regard of the other gender. We will call this third step an inquiry into the substantive comparability of the mutually exclusive programs provided to men and women.
Therefore, in this case we will examine a state-sponsored educational scheme offered by the Commonwealth of Virginia, under which the state provides a single-gender military-type college education to men and a single-gender college education with special leadership training to women, and determine (1) whether the state’s objective of providing single-gender education to its citizens may be considered a legitimate and important governmental objective; (2) whether the gender classification adopted is directly and substantially related to that purpose; and (3) whether the resulting mutual exclusion of women and men from each other’s institutions leaves open opportunities for those excluded to obtain substantively comparable benefits at their institution or through other means offered by the state. This is the special intermediate scrutiny test that we shall apply in deciding this case.*
rv
Turning to Virgima’s proposed VWIL program, we begin with the first part of the test and inquire into whether single-gender education constitutes a legitimate and important governmental objective, remembering that deference is to be accorded the state’s legislative will so long as the purpose is not pernicious and does not violate traditional notions of the role of government.
The provision of education is considered one of the most important functions of state and local government. See Brown v. Board of Educ., 347 U.S. 483, 493 (1954) (“Today, education is perhaps the most important function of state and local governments.”); Stroman v. Colleton County Sch. Dist., 981 F.2d 152, 158 (4th Cir.1992) (“Public education is recognized as one of the most important public services offered by state gov-*1238eminent.”). State and local governments routinely commit large portions of limited tax dollars to education and mandate that, through a certain educational level, attendance at school is legally required. Moreover, discussions of economic competitiveness and the root causes of social disorder commonly end in a discussion about the importance of public education.
As important as education is thought to be to the welfare of the people, it is nevertheless not a right secured to the people by the Constitution, see San Antonio School District v. Rodriguez, 411 U.S. 1, 35 (1973), and following from that reality, a citizen does not, in the absence of legislative will, have a right to demand a publicly financed education. Moreover, it is not the province of the courts to create such rights in the name of guaranteeing equal protection of the laws. Id. at 33.
When a state chooses to support college education, it need not provide all types of education, all disciplines, all methods, or all courses. A state with limited resources might, for example, subsidize a medical school without similarly subsidizing a law school, and its selection from among many permissible beneficial programs does not in and of itself constitute “unequal protection.” When a state narrows the range of its educational offerings, a narrowing of the class of those appropriately benefited is a necessary consequence. But a state may not establish a classification for admission, for example, to a medical school unrelated to its purpose of providing a medical education. Thus, we would expect judicial approval of an admissions classification based on intellectual capability, but not on race or national origin. In a similar vein, if a state were to choose to subsidize a conservatory of music, sponsoring-such a benefit might not fall beyond the range of what constitutes a legitimate and important state purpose. Yet those qualified to attend would naturally be a narrowly-drawn class of persons who qualified as the best musicians. Such a classification necessarily excludes most people, yet it is substantially related to the governmental objective. On the other hand, a classification for that conservatory defined by gender would probably not qualify as an appropriate classification.
Turning to this case, providing the option of a single-gender college education may be considered a legitimate and important aspect of a public system of higher education. That single-gender education at the college level is beneficial to both sexes is a fact established in this case. See United States v. Commonwealth of Virginia, 766 F.Supp. 1407, 1411—12 (W.D.Va.1991). Indeed, the briefs submitted in this case by the parties and amici curiae list a multitude of professional articles describing the benefits of single-gender education, especially for late adolescents coming out of high school. This should not be surprising in light of common experience that a sex-neutral atmosphere can be less distracting to late adolescents in an educational setting where the focus is properly on matters other than relationships between the sexes. Moreover, it is not surprising that the public, increasingly seeking admission to single-gender colleges, finds this objective to be important. A recent edition of a national magazine, devoted to an annual collection of statistical data and ratings about colleges and universities, reports:
After two decades as also-rans in higher education’s rush to embrace coeducation, women’s colleges are experiencing an unanticipated surge in enrollments and positive public attention.
# * # * * *
While the disproportionate distinction achieved by women who are alumnae of single-sex institutions ... is partially responsible for the enrollment boom, there are other explanations. Many attribute the newfound popularity to studies showing that girls in adolescence and beyond typically react to coeducational classrooms with “learned silence” and lowered aspirations. Others cite the diminished attraction of coeducation because of worrisome statistics on drinking and concerns about date rape and other violent crimes at institutions with both men and women, as well as the mounting criticism of large universities for seeming indifference to the quality of undergraduate education.
*1239“A Burst of Popularity,” U.S. News & World Report, Sept. 26, 1994.
Just as a state’s provision of publicly financed education to its citizens is a legitimate and important governmental objective, so too is a state’s opting for single-gender education as one particular pedagogical technique among many. Although there remains some disagreement among the experts about the extent of the benefits of single-gender education, it is not our role to resolve that issue. It is enough that there is a growing consensus in the professional community that a sexually homogeneous environment yields concrete educational benefits. Thus, we should defer to a state’s selection of educational techniques when we conclude, as we do here, that the purpose of providing single-gender education is not pernicious and falls within the range of the traditional governmental objective of providing citizens higher education. Accordingly, we conclude that Virginia has met the first part of our intermediate scrutiny test.
V
When applying the special intermediate scrutiny test for classifications based on homogeneity of gender in the context of higher education, we next consider whether that classification is substantially related to the state’s purpose. When combined with the third part of the test, ie., the inquiry into whether excluded men and women have opportunities to obtain substantively comparable benefits, this inquiry scrutinizes the means by which the state chooses to obtain its objective.
Single-gender education provides an educational environment in which the student population is of one sex, providing the assumed benefit that those students are not distracted by the presence of the other sex. Even though it may be offered to both genders through separate institutions, separate campuses, or even separate classrooms, a single-gender educational program necessarily excludes members of the gender not included in that institution, campus, or classroom. The importance of the classification is not the fact that the student body is male or female, but that it is of the same gender, whichever is chosen for the particular program. But the only way to realize the benefits of homogeneity of gender is to limit admission to one gender. Thus, the means of classifying by gender are focused on the single-gender educational purpose as directly as the nature of the objective allows.
The classification for single-gender education at VMI is also directly related to achieving the results of an adversative method in a military environment. The adversa-tive method was not designed to exclude women, but seized on the possibility, in a sexually homogeneous environment, of grating egos and setting the aggressiveness of one person against another through conflict, egalitarianism, lack of privacy, and stress— both physical and mental. The adversative method is intended to break down individualism and to instill the uniform values espoused by the institution. The methodology described, however, has never been tolerated in a sexually heterogeneous environment; indeed, we condemn it for good reason. If we were to place men and women into the adver-sative relationship inherent in the VMI program, we would destroy, at least for that period of the adversative training, any sense of decency that still permeates the relationship between the sexes.
Accordingly, to preserve the benefits of single-gender education, which Virginia has chosen to attain through separate institutions, the programs at VMI and Mary Baldwin College would of necessity exclude persons of the opposite gender, men at Mary Baldwin College and women at VMI. No more direct means could be adopted to accomplish the state’s objective of providing single-gender education at the institutional level. It is inherent in the benefit that men must be excluded from the women’s program and women from the men’s.
While we are satisfied that a classification for homogeneity of gender is necessary to provide single-gender education, at whatever level of separation, we must nevertheless, under the special intermediate scrutiny test that we are applying for such classification, be satisfied that both excluded men and excluded women have reasonable opportunities *1240to obtain benefits substantively comparable to those they are denied. That brings us to the final inquiry of this intermediate scrutiny test.
VI
In determining the substantive comparability of benefits, we are faced with at least two questions: how are the benefits from which one gender is excluded to be defined, and on what level and to what degree must other benefits be comparable.
The United States notes that VMI affords a unique type of military training as part of its educational program which cannot be duplicated in another institution. Even though it acknowledges that a parallel program could theoretically satisfy the requirements of the Equal Protection Clause, the United States argues that any such program must be identical to that of VMI. Because that cannot be accomplished, it concludes that women could only enjoy the unique benefits of the VMI program if VMI admits women.
The failure of the government’s syllogism, however, comes from its failing to follow its logic to completion. If we ordered VMI to admit women, the program would be irrevocably altered, forever denying its unique methodology to both women and men. Changes would have to be made to the ad-versative method, to the absence of privacy, and to the physical requirements of the program, all of which are part of VMI’s unique methodology. Certainly military training could be provided for women at VMI, but it would be substantially different from the training VMI cadets currently receive and would be closer to the programs offered by the U.S. military academies, which are already open to women. Thus, neither gender would experience the unique type of adversa-tive military training now utilized at VMI if VMI were to become coeducational.
Moreover, the government’s argument that a comparable opportunity requires an identical program is not sustained by the Equal Protection Clause. The advocation that laws require equal methods and equal results for different classes of people can no more be supported than the suggestion that two programs for two different classes of people can ever be identical. See Reed, 404 U.S. at 75; Jenness v. Fortson, 403 U.S. 431, 442 (1971) (“Sometimes the grossest discrimination can lie in treating things that are different as though they were exactly alike.”). And the alternative of allowing a state to provide benefits only when they could be provided in identical form to all of its citizens, regardless of whether they are similarly circumstanced, is justified only by a needless, and indeed baseless, demand for conformity.
Thus, if the state desires to offer the benefits of single-gender education to its citizens, the state must mitigate the effects of the resulting gender classification by affording to both genders benefits comparable in substance, but not . in form and detail.
VMI offers a publicly subsidized college education in a single-gender environment, resulting in a bachelor’s degree and intended to produce disciplined men of honor who are well-suited for leadership. Its method involves the use of the traditional classroom in a pervasive military environment. The VWIL program at Mary Baldwin College would likewise intend to provide an educational opportunity in a single-gender environment, leading to a bachelor’s degree coupled with discrete training designed specifically to prepare women for leadership. In considering the level of detail for any comparison of the two programs, we must, to achieve a meaningful comparison of substance, do more than simply recognize that both programs provide higher education leading to an undergraduate degree. But we should not reject programs that are aimed at achieving similar results, not generally available from other institutions of higher learning, simply because they differ in approach. In this case, both VMI and VWIL are focused on results beyond simply awarding an undergraduate degree. Both seek to teach discipline and prepare students for leadership. The missions are similar and the goals are the same. The mechanism for achieving the goals differ — VMI utilizing an adversa-tive and pervasive military regimen and VWIL proposing to utilize a structured environment reinforced by some military training and a concentration on leadership develop*1241ment — but the difference is attributable to a professional judgment of how best to provide the same opportunity.
To argue whether the adversative and pervasive military method applied to men at VMI should be applied to women at VWIL to reach better the goal of taught discipline and leadership in women makes for a rigorous debate among professional educators. The possibility of adapting the adversative methodology to women, setting woman against woman with the intended purpose of breaking individual spirit and instilling values, could succeed only if it is true that women, subjected to the same grating of mind and body, respond in the same way men do, and only then if a sufficient number of women necessary to make such a program work desired to participate in the program. Educational experts for the Commonwealth testified that women may not respond similarly and that if the state were to establish a women’s VMI-type program, the program would attract an insufficient number of participants to make the program work. The United States did not offer sufficient evidence to lead us to conclude that the Commonwealth’s expert testimony was clearly erroneous in this regard. But we need not resolve such details of methodology. This is the type of ongoing debate that is to be expected among substantively comparable institutions, and it reveals a vitality of professional concern which can lead to institutional betterment through adjustments down the road. In this case, the mission and goals are the same, and the methodologies for attaining the goals, while different, nevertheless are reasonably calculated to succeed at each institution. Those differences that do exist do not require that the important state purpose of providing single-gender education for both sexes be defeated in this case.
It is true that VWIL is at its incipiency, and the VWIL degree from Mary Baldwin College lacks the historical benefit and prestige of a degree from VMI. But such intangible benefits can never be created on command — they must be the byproduct of a longer-term effort. Moreover, to some extent, we compensate for this deficiency in the remedy section, below. For purposes of the Equal Protection Clause, however, we are satisfied that the programs to be offered at both institutions can be substantively comparable if VWIL is undertaken with a persistently high level of commitment by Virginia and that men and women mutually excluded by the two programs will not be denied the opportunity for an undergraduate education with discipline and special training in leadership. It is noteworthy that men and women are not limited to the choices available at these two institutions. Virginia provides a much broader array of opportunities in higher education through other state supported colleges and universities, including the coeducational military program at Virginia Polytechnic Institute and State University. See VMI I, 976 F.2d at 893 n. 1 & 898 n. 8.
In this case, we conclude that if the conditions that we impose below are fulfilled, the opportunities that would be open both to men and women are sufficiently comparable. We therefore are satisfied that the special intermediate scrutiny test defined for this case has been met, insofar as a proposed program can meet this test, by the VWIL program proposed at Mary Baldwin College.
VII
Were Virginia now building its higher educational program from the ground up and, as part of it, offering bachelor’s programs (1) at a male-only institution featuring a highly disciplined military environment, (2) at a female-only institution featuring a highly disciplined leadership program in a non-military environment, and (3) at a third institution offering a broad array of subjects and methods in a coeducational environment, our analysis would end here with approval of the program against an equal protection challenge.
In this case, however, there is an added element created by the presence now of VMI as an ongoing and successful institution with a long history and the absence now of a comparable single-gender women’s institution. Virginia’s proposal for Mary Baldwin College is just that — a proposal. Virginia has undertaken what appears to be a serious effort at developing a plan to meet this historic deficiency. Virginia appointed a task *1242force of professionals to design a new program, designed a program aimed at special leadership for women, and funded the proposed program at the same per capita levels at which it funds VMI. In addition, governmental officials in Virginia seem to be supporting the new program at every level. In our earlier opinion we noted some ambivalence in that regard. Then-Governor Douglas Wilder had favored coeducation at VMI in the face of no other alternative, and state education officials favored a separate program, or some other course, leading the state, as a party, to bow out of the liability phase of the litigation as a house divided. Governor Wilder is now firmly behind the VWIL program as is current Governor George Allen. Moreover, the Virginia legislature has supported the program by providing what appears to be adequate funding and by promising to increase the level of funding, should the response require it.
Nevertheless, a state’s response to a court ordered correction of a Fourteenth Amendment violation is given under command and therefore must be viewed with some skepticism. While the court was assured at oral argument that the program proposed at Mary Baldwin College was serious and had the full support of the state, the important question remains whether Virginia will implement the program with the intensity and perseverance necessary to provide a substantively comparable opportunity for women, so that when VWIL is established we will not conclude that the value of the benefits provided by that program, when compared to VMI, tends “to lessen the dignity, respect, or societal regard” of women. To allay any skepticism and assure eradication of the constitutional violation, we therefore find it essential, during the early stage of VWIL’s history, to be assured affirmatively that a high level of state support continues.
Accordingly, while we affirm the judgment of the district court, which has issued an injunction mandating implementation of the plan and retaining jurisdiction to oversee the implementation, we are remanding the case ■with instructions that the court include, as part of its oversight of the plan’s implementation, a specific review to ensure that (1) the program is headed by a well-qualified, motivated administrator, attracted by a level of compensation suited for the position; (2) the program is well-promoted to potentially qualified candidates; (3) the program includes a commitment for adequate funding by the state for the near term; and (4) the program includes.a mechanism for continuing review by qualified professional educators so that its elements may be adjusted as necessary to keep the program aimed not only at providing a quality bachelor’s degree but also at affording the additional element of taught discipline and leadership training for women.
AFFIRMED AND REMANDED.
The dissenting opinion has improperly characterized this test as one for "allowable separate-but-equal state-supported educational institutions.” This misunderstands the standard we utilize.
As a general principle of equal protection jurisprudence, when there is no meaningful and relevant difference between two classes of persons for purposes of a given state regulation, equality is demanded, and "separate but equal” does not fulfill the demand. When there is a difference between two classes of persons, then separate and different facilities for each class may satisfy equal protection if the difference in facilities is sufficiently related to the nature of the difference between the classes.
In this case, we do not espouse a “separate-but-equal” test and never discuss "separate-but-equal facilities.” Rather, the test we utilize would allow separate and substantively comparable facilities where a state justifies its offering of single gender education as a legitimate governmental objective.