with whom
Justice Blackmun and Justice Stevens join, and with whom Justice O’Connor joins as to Part II, dissenting.Because I conclude that Kentucky’s provision of different procedures for the institutionalization of the mentally retarded and the mentally ill is not supported by any rational justification, I respectfully dissent.
I
To begin with, the Court declines to address Doe’s argument that we should employ strict or heightened scrutiny in assessing the disparity of treatment challenged here.1 *336While I may disagree with the Court’s basis for its conclusion that this argument is not “properly presented,” ante, at 319, I too would decline to address the contention that strict or heightened scrutiny applies. I conclude that the distinctions wrought by the Kentucky scheme cannot survive even that rational-basis scrutiny, requiring a rational relationship *337between the disparity of treatment and some legitimate governmental purpose, which we have previously applied to a classification on the basis of mental disability, see Cleburne v. Cleburne Living Center, Inc., 473 U. S. 432, 446-447 (1985), and therefore I need not reach the question of whether scrutiny more searching than Cleburne’s should be applied.2 Cleburne was the most recent instance in which we addressed a classification on the basis of mental disability, as we did by enquiring into record support for the State’s proffered justifications, and examining the distinction in treatment in light of the purposes put forward to support it. See id., at 450. While the Court cites Cleburne once, and does not purport to overrule it, neither does the Court apply it, and at the end of the day Cleburne’s status is left uncertain. I would follow Cleburne here.
II
Obviously there are differences between mental retardation and mental illness. They are distinct conditions, they have different manifestations, they require different forms of care or treatment, and the course of each differs. It is without doubt permissible for the State to treat those who are mentally retarded differently in some respects from those who are mentally ill. The question here, however, is whether some difference between the two conditions ration*338ally can justify the particular disparate treatment accorded under this Kentucky statute.
The first distinction wrought by the statute is the imposition of a lesser standard of proof for involuntary institutionalization where the alleged basis of a need for confinement is mental retardation rather than mental illness. As the Court observes, four specific propositions must be proven before a person may be involuntarily institutionalized on the basis of mental retardation: “that: (1) [t]he person is a mentally retarded person; (2) [t]he person presents a danger or a threat of danger to self, family, or others; (3) [t]he least restrictive alternative mode of treatment presently available requires placement in [a state-run institution]; and (4) [treatment that can reasonably benefit the person is available in [a state-run institution].” Ky. Rev. Stat. Ann. §202B.040 (Michie 1991). At issue in this case is only the application of this provision to adults who have not been shown to be mentally retarded, but who are simply alleged to be. The subject of such a proceeding retains as full an interest in liberty as anyone else. The State of Kentucky has deemed this liberty interest so precious that, before one may be institutionalized on the basis of mental illness, the statutory prerequisites must be shown “beyond a reasonable doubt.” § 202A.076(2).3 However, when the allegation against the individual is one of mental retardation, he is deprived of the protection of that high burden of proof. The first question here, then, is whether, in light of the State’s decision to provide that high burden of proof in involuntary commitment *339proceedings where illness is alleged, there is something about mental retardation that can rationally justify provision of less protection.
In upholding this disparate treatment, the Court relies first on the State’s assertion that mental retardation is easier to diagnose than mental illness. It concludes that the discrimination in burdens of proof is rational because the lessened “ ‘risk of error’ ” resulting from the higher burden of proof, see ante, at 322 (quoting Addington v. Texas, 441 U. S. 418, 423 (1979)), can be understood to offset a greater “ris[k] of an erroneous determination that the subject of a commitment proceeding has the condition in question” when the allegation is one of mental illness rather than mental retardation, ante, at 322. The Court reaches essentially the same conclusion with respect to the second prerequisite, that the individual present a danger or threat of danger to himself or others. See ante, at 324 (a determination of dangerousness may be made with “more accura[cy]” with respect to the mentally retarded than the mentally ill).
In concluding, however, that the demands of minimal rationality are satisfied if burdens of proof rise simply with difficulties of proof, the Court misunderstands the principal object in setting burdens. It is no coincidence that difficult issues in civil cases are not subject to proof beyond a reasonable doubt and that even the most garden variety elements in criminal cases are not to be satisfied by a preponderance of evidence. The reason for this is that burdens of proof are assigned and risks of error are allocated not to reflect the mere difficulty of avoiding error, but the importance of avoiding it as judged after a thorough consideration of those respective interests of the parties that will be affected by the allocation. See Addington, 441 U. S., at 425.
In a civil commitment proceeding, on the State’s side of the balance, are the interests of protecting society from those posing dangers and protecting the ill or helpless individual from his own incapacities. Id., at 426. On the other *340side, it is clear that “[i]n cases involving individual rights, whether criminal or civil, ‘[t]he standard of proof [at a minimum] reflects the value society places on individual liberty,’ ” id., at 425 (brackets in original and citation omitted), which encompasses both freedom from restraint and freedom from the stigma that restraint and its justifications impose on an institutionalized person, id., at 425-426.
The question whether a lower burden of proof is rationally justified, then, turns not only on whether ease of diagnosis and proof of dangerousness differ as between cases of illness and retardation, but also on whether there are differences in the respective interests of the public and the subjects of the commitment proceedings, such that the two groups subject to commitment can rationally be treated differently by imposing a lower standard of proof for commitment of the retarded.4 The answer is clearly that they cannot. While difficulty of proof, and of interpretation of evidence, could legitimately counsel against setting the standard so high that the State may be unable to satisfy it (thereby effectively thwarting efforts to satisfy legitimate interests in protection, care, and treatment), see id., at 429, that would at most justify a lower standard in the allegedly more difficult cases of illness, not in the easier cases of retardation. We do not lower burdens of proof merely because it is easy to prove the proposition at issue, nor do we raise them merely because it is difficult.5 Nor do any other reasonably conceivable facts *341cut in favor of the distinction in treatment drawn by the Kentucky statute. Both the ill and the retarded may be dangerous, each may require care, and the State’s interest is seemingly of equal strength in each category of cases. No one has or would argue that the value of liberty varies somehow depending on whether one is alleged to be ill or retarded, and a mentally retarded person has as much to lose by civil commitment to an institution as a mentally ill counterpart, including loss of liberty to “choosfe] his own friends and companions, selec[t] daily activities, decid[e] what to eat, and retai[n] a level of personal privacy,” among other things. Brief for American Association on Mental Retardation (AAMR) et al. as Amici Curiae 12 (AAMR Br.). We do not presume that a curtailment of the liberty of those who are disabled is, because of their disability, less severe than the same loss to those who are ill. Even if the individuals subject to involuntary commitment proceedings previously had been shown to be mentally retarded, they would thus still retain their “strong,” legally cognizable interest in their liberty. Cf. Foucha, 504 U. S., at 88 (O’Connor, J., concurring in part and concurring in judgment). Even assuming, then, that the assertion of different degrees of difficulty of proof both of mental illness and mental retardation and of the dangerousness inherent in each condition is true (an assertion for which there is no support in the record), it lends not a shred of rational support to the decision to discriminate against the retarded in allocating the risk of erroneous curtailment of liberty.
The Court also rests its conclusion on the view that “it would have been plausible for the Kentucky Legislature to believe that most mentally retarded individuals who are *342committed receive treatment that is . . . less invasive tha[n] that to which the mentally ill are subjected.” Ante, at 326. Nothing cited by the Court, however, demonstrates that such a belief would have been plausible for the Kentucky Legislature, nor does the Court's discussion render it plausible now. Cf. United States Railroad Retirement Bd. v. Fritz, 449 U. S. 166, 179 (1980) (under rational-basis scrutiny disparate treatment must be justified by “plausible reasons”). One example of the invasiveness to which the Court refers is the use of (and the results of the administration of) psychotropic drugs. I take no exception to the proposition that they are extensively used in treating mental illness. See ante, at 325 (citing authorities for the proposition that drugs are used in treating mental illness). Nor do I except to the proposition that the appropriate and perhaps characteristic response to mental retardation, but not to mental illness, is that kind of training in the necessities of self-sufficiency known as “habilitation.” See ibid, (citing authorities describing such training).
Neither of these propositions tells us, however, that the same invasive mind-altering medication prescribed for mental illness is not also used in responding to mental retardation. And in fact, any apparent plausibility in the Court's suggestion that “the mentally retarded in general are not subjected to th[is] medical treatment],” ibid., dissipates the moment we examine readily available material on the subject, including studies of institutional practices affecting the retarded comparable to those studies concerning the treatment of mental illness cited by the Court. One recent examination of institutions for the mentally retarded in Kentucky's neighboring State of Missouri, for example, found that 76% of the institutionalized retarded receive some type of psychoactive drug and that fully 54% receive psychotropic drugs. See Intagliata & Rinck, Psychoactive Drug Use in Public and Community Residential Facilities for Mentally Retardéd Persons, 21 Psychopharmacology Bull. 268, 272-*343273 (1985). Another study, this one national in scope, found that 38% of the residents of institutions for the mentally retarded receive psychotropic drugs. See Hill, Balow, & Bruininks, A National Study of Prescribed Drugs in Institutions and Community Residential Facilities for Mentally Retarded People, 21 Psychopharmacology Bull. 279, 283 (1985). “Surveys conducted within institutions [for the mentally retarded] have generally shown prevalences in the range of 30% to 50% of residents receiving psychotropic drugs at any given time.” Aman & Singh, Pharmacological Intervention, in Handbook of Mental Retardation 347, 348 (J. Matson & J. Mulick eds., 2d ed. 1991) (hereinafter Handbook of Mental Retardation).
Psychotropic drugs, according to the available material, are not only used to treat the institutionalized retarded, but are often misused. Indeed, the findings of fact by a United States District Court in North Carolina, another State nearby Kentucky, show that in three hospitals, 73% of persons committed as mentally retarded were receiving antipsychotic drugs. Less than half of these individuals had been diagnosed as mentally ill as well as mentally retarded following their commitment on the latter ground. See Thomas S. v. Flaherty, 699 F. Supp. 1178, 1187 (WDNC 1988), aff’d, 902 F. 2d 250 (CA4), cert. denied, 498 U. S. 951-952 (1990). The District Court found that the institutionalized retarded plaintiffs “have been seriously endangered and injured by the inappropriate use of antipsychotic drugs.” Flaherty, supra, at 1186. See also Halderman v. Pennhurst State School Hospital, 446 F. Supp. 1295, 1307-1308 (ED Pa. 1977) (discussing evidence that 51% of the residents of a state institution for the mentally retarded received psychotropic drugs though less than one-third of those who received the drugs were monitored to determine the effectiveness of the treatment), aff’d, 612 F. 2d 84 (CA3 1979), rev’d on other grounds, 451 U. S. 1 (1981); Bates, Smeltzer, & Arnoczky, Appropriate and Inappropriate Use of Psychotherapeutic *344Medications for Institutionalized Mentally Retarded Persons, 90 Am. J. Mental Deficiency 363 (1986) (finding that between 39% and 54% of medications prescribed to mentally retarded persons are inappropriate for the conditions diagnosed).
These facts are consistent with a law review study of drugs employed in treating retardation, which observed that the reduction in the need for institutional staff resulting from the use of sedating drugs has promoted drug use in responding to retardation despite “frightening adverse effects [including the suppression of] learning and intellectual development.” Plotkin & Gill, Invisible Manacles: Drugging Mentally Retarded People, 31 Stan. L. Rev. 637, 638 (1979). There being nothing in the record to suggest that Kentucky’s institutions are free from these practices, and no reason whatever to assume so, there simply is no plausible basis for the Court’s assumption that the institutional response to mental retardation is in the main less intrusive in this way than treatment of mental illness.
The Court also suggests that medical treatment for the mentally retarded is less invasive than in the case of the mentally ill because the mentally ill are subjected to psychiatric treatment that may involve intrusive enquiries into the patient’s innermost thoughts. See ante, at 324-325. Again, I do not disagree that the mentally ill are often subject to intrusive psychiatric therapy. But the mentally retarded too are subject to intrusive therapy, as the available material on the medical treatment of the mentally retarded demonstrates. The mentally retarded are often subjected to behavior modification therapy to correct, among other things, anxiety disorders, phobias, hyperactivity, and antisocial behavior, therapy that may include aversive conditioning as well as forced exposure to objects that trigger severe anxiety reactions. See McNally, Anxiety and Phobias, in Handbook of Mental Retardation 413-423; Mulick, Hammer, & Dura, Assessment and Management of Antisocial and Hyperactive Behavior, in *345Handbook of Mental Retardation 397-412; Gardner, Use of Behavior Therapy with the Mentally Retarded, in Psychiatric Approaches to Mental Retardation 250-275 (F. Menolascino ed. 1970). Like drug therapy, psychiatric therapy for the mentally retarded can be, and has been, misused. In one recent case, a Federal District Court found that “aversive procedures [including seclusion and physical restraints were] being inappropriately used with no evidence for their effectiveness and no relationship between the choice of the procedure and the analysis of the cause of the problem[,]... placing] clients at extreme risk for maltreatment.” Lelsz v. Kavanagh, 673 F. Supp. 828, 850 (ND Tex.) (internal quotation marks and citation omitted), rev’d on unrelated grounds, 824 F. 2d 372 (CA5 1987). Invasive behavior therapy for the mentally retarded, finally, is often employed together with drug therapy. See McNally, supra, at 413-423; Mulick, Hammer, & Dura, supra, at 397-412.
The same sorts of published authorities on which the Court relies, in sum, refiite the contention that “[t]he prevailing methods of treatment for the mentally retarded, as a general rule, are much less invasive than are those given the mentally ill.” Ante, at 324.6 The available literature indicates that psychotropic drugs and invasive therapy are routinely administered to the retarded as well as the mentally *346ill, and there are no apparent differences of therapeutic regimes that would plausibly explain less rigorous commitment standards for those alleged to be mentally retarded than for those alleged to be mentally ill.7
Ill
With respect to the involvement of family members and guardians in the commitment proceeding, the Court holds it to be justified by the fact that mental retardation “has its onset during a person’s developmental period,” while mental illness “may arise or manifest itself with suddenness only after minority.” Ante, at 329. The Court suggests that a mentally ill person’s parents may have “ceased to provide care and support” for him well before the onset of illness, whereas parents are more likely to have retained connection with a retarded son or daughter, whose “proper course of treatment” may depend on matters related to “observations made in a household setting.” Ibid.
These suggested distinctions, if true, would apparently not apply to guardians, whose legal obligations to protect the persons and estates of their wards would seem to require as much connection to the one class of people as to the other. *347In any event, although these differences might justify a scheme in which immediate relatives and guardians were automatically called as witnesses in cases seeking institutionalization on the basis of mental retardation,8 they are completely unrelated to those aspects of the statute to which Doe objects: permitting these immediate relatives and guardians to be involved “as parties” so as to give them, among other things, the right to appeal as “adverse” a decision not to institutionalize the individual who is subject to the proceedings. Where the third party supports commitment, someone who is alleged to be retarded is faced not only with a second advocate for institutionalization, but with a second prosecutor with the capacity to call and cross-examine witnesses, to obtain expert testimony and to raise an appeal that might not otherwise be taken, whereas a person said to require commitment on the basis of mental illness is not. This is no mere theoretical difference, and my suggestion that relatives or guardians may support curtailment of liberty finds support in the record in this case. It indicates that of the 431 commitments to Kentucky’s state-run institutions for the mentally retarded during a period between 1982 and the middle of 1985, all but one were achieved through the application or consent of family members or guardians. See Record, State’s Answers to Plaintiff’s First Set of Interrogatories 2,17.
The Court simply points to no characteristic of mental retardation that could rationally justify imposing this burden of a second prosecutor on those alleged to be mentally retarded where the State has decided not to impose it upon those alleged to be mentally ill. Even if we assumed a generally more regular connection between the relatives and guardians of those alleged to be retarded than those said to *348be mentally ill, it would not explain why the former should be subject to a second prosecutor when the latter are not.
The same may be said about the Court’s second suggested justification, that the mentally ill may have a need for privacy not shown by the retarded. Even assuming the ill need some additional privacy, and that participation of others in the commitment proceeding should therefore be limited “to the smallest group compatible with due process,” ante, at 329, why should the retarded be subject to a second prosecutor? The Court provides no answer.9
Without plausible justification, Kentucky is being allowed to draw a distinction that is difficult to see as resting on anything other than the stereotypical assumption that the retarded are “perpetual children,” an assumption that has historically been taken to justify the disrespect and “grotesque mistreatment” to which the retarded have been subjected. See Cleburne, 473 U. S., at 454 (Stevens, J., concurring) (internal quotation marks and citation omitted). As we said in Cleburne, the mentally retarded are not “all cut from the same pattern:... they range from those whose disability is not immediately evident to those who must be constantly cared for.” Id., at 442. In recent times, at least when imposing the responsibilities of citizenship, our jurisprudence has seemed to reject the analogy between mentally retarded adults and nondisabled children. See, e.g., Penry v. Lynaugh, 492 U. S. 302, 338 (1989) (controlling opinion of O’Connor, J.) (not “all mentally retarded people ...— by virtue of their mental retardation alone, and apart from any individualized consideration of their personal responsibility — inevitably lack the cognitive, volitional, and moral capacity to act with the degree of culpability associated with the death penalty”); see also id., at 340 (“reliance on mental *349age to measure the capabilities of a retarded person for purposes of the Eighth Amendment could have a disempowering effect if applied in other areas of the law”). But cf. ante, at 331 (citing Parham v. J. R., 442 U. S. 584 (1979), a case about parents’ rights over their minor children). When the State of Kentucky sets up its respective schemes for institutionalization on the basis of mental illness and mental retardation, it too is obliged to reject that analogy, and to rest any difference in standards for involuntary commitment as between the ill and the retarded on some plausible reason.
IV
In the absence of any rational justification for the disparate treatment here either with respect to the burdens of proof or the participation of third parties in institutionalization proceedings, I would affirm the judgment of the Court of Appeals. Because of my conclusion, that the statute violates equal protection, I do not reach the question of its validity under the Due Process Clause.
Doe relies, first, on the nature of the right at stake, citing our decision last Term in Foucha v. Louisiana, 504 U. S. 71 (1992). There we were faced with an equal protection challenge to a Louisiana statute authorizing continued commitment of currently sane insanity acquittees under standards that were not applied to criminal convicts who had completed their prison terms or were about to do so. The insanity acquittee was kept incarcerated in a mental institution unless he could prove he was not dangerous, see La. Code Crim. Proc. Ann., Art. 657 (West Supp. 1993), whereas “Louisiana law,” as Justice White wrote, did “not provide for similar confinement for other classes of persons who have committed criminal acts and who cannot later prove they would not be dangerous. Criminals who have completed their prison terms, or are about to do so, are an obvious and large category of such persons .... However, state law does not allow for th[e] continuing confinement [of criminals who may be unable to prove they would not be dangerous] based merely on dangerousness. ... Freedom from physical restraint being a fundamental right, the State must have a particularly convincing reason, which it has not put forward, *336for such discrimination against insanity acquittees who are no longer mentally ill.” Foucha, 504 U. S., at 85-86 (plurality opinion of White, J., joined by Blackmun, Stevens, and Souter, JJ.); see also id., at 88 (O’Connor, J., concurring in part and concurring in judgment) (“Although I think it unnecessary to reach equal protection issues on the facts before us, the permissibility of holding an acquittee who is not mentally ill longer than a person convicted of the same crimes could be imprisoned is open to serious question”). Because of the "'massive curtailment of liberty’” undoubtedly involved in involuntary civil commitment and institutionalization, see Vitek v. Jones, 445 U. S. 480, 491 (1980) (quoting Humphrey v. Cady, 405 U. S. 504, 509 (1972)), Doe argues that heightened scrutiny applies under Foucha when those alleged to be mentally retarded are denied the protection afforded another “obvious and large category” of potential civil committees, those said to be mentally ill.
Doe also argues that the discrimination here has a second aspect that justifies application of strict or heightened scrutiny, in its classification on the basis of mental retardation. Although he recognizes that this Court held in 1985 that retarded individuals are not a quasi-suspect class, see Cleburne v. Cleburne Living Center, Inc., 473 U. S. 432, 442-447 (1985), he argues that the subsequently enacted Americans With Disabilities Act of 1990 (ADA) amounts to an exercise of Congress’s power under § 5 of the Fourteenth Amendment to secure the guarantees of the Equal Protection Clause to the disabled. See Katzenbach v. Morgan, 3S4 U. S. 641, 651 (1966). The ADA includes findings that people with disabilities (among whom are included those with mental impairments that Doe argues include mental retardation, see 42 U. S. C. § 12102(2)(A) (1988 ed., Supp. Ill)) “are a discrete and insular minority who have been faced with restrictions and limitations, subjected to a history of purposeful unequal treatment, and relegated to a position of political powerlessness in our society .. ..” § 12101(a)(7). Doe argues that this and other findings, together with expressions of purpose contained in the ADA, amount to a clear indication from Congress “that all individuals with disabilities, including individuals with mental retardation should be treated as a suspect class.” Brief for Respondents 29-30.
This approach complies with “two of the cardinal rules governing the federal courts: one, never to anticipate a question of constitutional law in advance of the necessity of deciding it; the other never to formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied,” Brockett v. Spokane Arcades, Inc., 472 U. S. 491, 501 (1985) (citations, internal quotation marks, and brackets omitted), and is consistent with our past practice. See, e. g., Hooper v. Bernalillo County Assessor, 472 U. S. 612, 618 (1985) (declining to decide whether to apply heightened scrutiny where classification failed rational-basis test); cf. Mississippi Univ. for Women v. Hogan, 458 U. S. 718, 724, n. 9 (1982) (declining to decide whether to apply strict scrutiny where classification could not survive heightened scrutiny).
As the Court notes, the statutory prerequisites are substantially identical for commitment on the basis of illness and retardation. Commitment on the ground of mental illness requires proof beyond a reasonable doubt that an individual “is a mentally ill person: (1) [w]ho presents a danger or threat of danger to self, family or others as a result of the mental illness; (2) [w]ho can reasonably benefit from treatment; and (3) [f]or whom hospitalization is the least restrictive alternative mode of treatment presently available.” Ky. Rev. Stat. Ann. §202A.026 (Michie 1991).
In addition to the two prerequisites mentioned in the text, the State must also prove that commitment would be beneficial and the least restrictive alternative method of treatment. The Court does not contend that there is any rational justification for imposition of a lowered burden of proof with respect to these prerequisites for institutionalization in those cases where the allegation is one of retardation and not illness. See ante, at 324.
And indeed, to the extent Addington v. Texas, 441 U. S. 418 (1979), does discuss the difficulty of diagnosing mental illness, see id., at 429-430, it supports use only of a lesser standard of proof because of the practical problems created by a supposed “serious question as to whether a state *341could ever prove beyond a reasonable doubt that an individual is both mentally ill and likely to be dangerous,” id., at 429. Of course, in this case Kentucky has determined that the liberty of those alleged to be mentally ill is sufficiently precious that the State should assume the risk inherent in use of that higher standard.
I also see little point in the Court’s excursion into the historical difference in treatment between so-called “idiots,” and so-called “lunatics.” See ante, at 326. Surely the Court does not intend to suggest that the irrational and scientifically unsupported beliefs of pre-19th-century England can support any distinction in treatment between the mentally ill and the mentally retarded today. At that time, “lunatics” were “[s]een as demonically possessed or the products of parental sin [and] were often punished or left to perish.” See S. Herr, Rights and Advocacy for Retarded People 9 (1983). The primary purpose of an adjudication of “idiocy” appears to have been to “depriv[e] [an individual] of [his] property and its profits.” Id,., at. 10. Those without wealth “were dealt with like other destitute or vagrant persons through workhouses and houses of correction.” Id,., at 11.
Petitioner also argues that mental retardation is different from most cases of mental illness in being a permanent condition that may require a lifetime of care. See Brief for Petitioner 31. But petitioner completely fails to explain how the permanence of the condition or the likely need of lifetime care can rationally justify a regime in which those alleged to require institutionalization based on mental retardation face a greater risk of erroneous curtailment of liberty than those who are alleged to require it based on mental illness. The distinction proffered by the State (accepting it to be factually accurate and not based merely on stereotype) cuts quite the other way. The possibility that a condition once thought to justify commitment will last a lifetime suggests that a person committed to an institution on the basis of mental retardation is less likely to regain his liberty than one institutionalized on some other basis. If this could rationally justify any disparity in commitment standards, it could only be in requiring stricter protection in mental retardation cases than in those based on mental illness, not the other way around.
Of course both guardians and relatives can already act as witnesses in each kind of proceeding subject only to the limitations of relevance and interest.
I also note that the Court provides no support for its speculation that an adult who develops mental illness will have a greater need or desire for privacy in an involuntary commitment proceeding than an adult who is mentally retarded.