delivered the opinion of the Court.

The question presented is whether petitioner, who was committed to a mental hospital upon being acquitted of a criminal offense by reason of insanity, must be released because he has been hospitalized for a period longer than he might have served in prison had he been convicted.

In the District of Columbia a criminal defendant may be acquitted by reason of insanity if his insanity is “affirmatively established by a preponderance of the evidence.” D. C. Code § 24 — 801(j) (1981).1 If he successfully invokes the insanity defense, he is committed to a mental hospital. §24-301(d)(l).2 The statute provides several ways of ob*357taining release. Within 50 days of commitment the acquittee is entitled to a judicial hearing to determine his eligibility for release, at which he has the burden of proving by a preponderance of the evidence that he is no longer mentally ill or dangerous. § 24 — 301(d)(2).3 If he fails to meet this burden at the 50-day hearing, the committed acquittee subsequently may be released, with court approval, upon certification of his recovery by the hospital chief of service. §24-301(e).4 *358Alternatively, the acquittee is entitled to a judicial hearing every six months at which he may establish by a preponderance of the evidence that he is entitled to release. §24-301(k).5

Independent of its provision for the commitment of insanity acquittees, the District of Columbia also has adopted a civil-commitment procedure, under which an individual may be committed upon clear and convincing proof by the Govern*359ment that he is mentally ill and likely to injure himself or others. § 21-545(b).6 The individual may demand a jury in the civil-commitment proceeding. § 21-544. Once committed, a patient may be released at any time upon certification of recovery by the hospital chief of service. §§ 21-546, 21-548. Alternatively, the patient is entitled after the first 90 days, and subsequently at 6-month intervals, to request a judicial hearing at which he may gain his release by proving by a preponderance of the evidence that he is no longer mentally ill or dangerous. §§21-546, 21-547; see Dixon v. Jacobs, 138 U. S. App. D. C. 319, 328, 427 F. 2d 589, 598 (1970).

H-Í

On September 19, 1975, petitioner was arrested for attempting to steal a jacket from a department store. The next day he was arraigned in the District of Columbia Superior Court on a charge of attempted petit larceny, a misdemeanor punishable by a maximum prison sentence of one year. §§22-103, 22-2202. The court ordered petitioner committed to St. Elizabeths, a public hospital for the mentally ill, for a determination of his competency to stand trial.7 On March 1, 1976, a hospital psychologist submitted a report to the court stating that petitioner was competent to stand trial, that petitioner suffered from “Schizophrenia, paranoid *360type,” and that petitioner’s alleged offense was “the product of his mental disease.” Record 51. The court ruled that petitioner was competent to stand trial. Petitioner subsequently decided to plead not guilty by reason of insanity. The Government did not contest the plea, and it entered into a stipulation of facts with petitioner. On March 12,1976, the Superior Court found petitioner not guilty by reason of insanity and committed him to St. Elizabeths pursuant to § 24-301(d)(l).

On May 25, 1976, the court held the 50-day hearing required by § 24-301(d)(2)(A). A psychologist from St. Eliza-beths testified on behalf of the Government that, in the opinion of the staff, petitioner continued to suffer from paranoid schizophrenia and that “because his illness is still quite active, he is still a danger to himself and to others.” Tr. 9. Petitioner’s counsel conducted a brief cross-examination, and presented no evidence.8 The court then found that “the defendant-patient is mentally ill and as a result of his mental illness, at this time, he constitutes a danger to himself or others.” Id., at 13. Petitioner was returned to St. Elizabeths. Petitioner obtained new counsel and, following some procedural confusion, a second release hearing was held on February 22, 1977. By that date petitioner had been hospitalized for more than one year, the maximum period he could have spent in prison if he had been convicted. On that basis he demanded that he be released unconditionally or recommitted pursuant to the civil-commitment standards in § 21— 545(b), including a jury trial and proof by clear and convincing evidence of his mental illness and dangerousness. The Superior Court denied petitioner’s request for a civil-commitment hearing, reaffirmed the findings made at the *361May 25, 1976, hearing, and continued petitioner’s commitment to St. Elizabeths.9

Petitioner appealed to the District of Columbia Court of Appeals. A panel of the court affirmed the Superior Court, 396 A. 2d 183 (1978), but then granted rehearing and reversed, 411 A. 2d 624 (1980). Finally, the court heard the case en banc and affirmed the judgment of the Superior Court. 432 A. 2d 364 (1981). The Court of Appeals rejected the argument “that the length of the prison sentence [petitioner] might have received determines when he is entitled to release or civil commitment under Title 24 of the D. C. Code.” Id., at 368. It then held that the various statutory differences between civil commitment and commitment of insanity acquittees were justified under the equal protection component of the Fifth Amendment. Id., at 371-376.

We granted certiorari, 454 U. S. 1141 (1982), and now affirm.

Ill

It is clear that “commitment for any purpose constitutes a significant deprivation of liberty that requires due process protection.” Addington v. Texas, 441 U. S. 418, 425 (1979). Therefore, a State must have “a constitutionally adequate purpose for the confinement.” O’Connor v. Donaldson, 422 U. S. 563, 574 (1975). Congress has determined that a criminal defendant found not guilty by reason of insanity in the District of Columbia should be committed indefinitely to a mental institution for treatment and the protection of society. See H. R. Rep. No. 91-907, pp. 73-74 (1970); 432 A. 2d, at 371 (“[T]he District of Columbia statutory scheme for com*362mitment of insane criminals is ... a regulatory, prophylactic statute, based on a legitimate governmental interest in protecting society and rehabilitating mental patients”). Petitioner does not contest the Government’s authority to commit a mentally ill and dangerous person indefinitely to a mental institution, but rather contends that “the petitioner’s trial was not a constitutionally adequate hearing to justify an indefinite commitment.” Brief for Petitioner 14.

Petitioner’s argument rests principally on Addington v. Texas, supra, in which the Court held that the Due Process Clause requires the State in a civil-commitment proceeding to demonstrate by clear and convincing evidence that the individual is mentally ill and dangerous. 441 U. S., at 426-427. Petitioner contends that these due process standards were not met in his case because the judgment of not guilty by reason of insanity did not constitute a finding of present mental illness and dangerousness and because it was established only by a preponderance of the evidence.10 Peti*363tioner then concludes that the Government’s only conceivably legitimate justification for automatic commitment is to ensure that insanity acquittees do not escape confinement entirely, and that this interest can justify commitment at most for a period equal to the maximum prison sentence the acquit-tee could have received if convicted. Because petitioner has been hospitalized for longer than the one year he might have served in prison, he asserts that he should be released unconditionally or recommitted under the District’s civil-commitment procedures.11

A

. We turn first to the question whether the finding of insanity at the criminal trial is sufficiently probative of mental illness and dangerousness to justify commitment. A verdict of not guilty by reason of insanity establishes two facts: (i) the defendant committed an act that constitutes a criminal offense, and (ii) he committed the act because of mental illness. *364Congress has determined that these findings constitute an adequate basis for hospitalizing the acquittee as a dangerous and mentally ill person. See H. R. Rep. No. 91-907, supra, at 74 (expressing fear that “dangerous criminals, particularly psychopaths, [may] win acquittals of serious criminal charges on grounds of insanity” and yet “escape hospital commitment”); S. Rep. No. 1170, 84th Cong., 1st Sess., 13 (1955) (“Where [the] accused has pleaded insanity as a defense to a crime, and the jury has found that the defendant was, in fact, insane at the time the crime was committed, it is just and reasonable in the Committee’s opinion that the insanity, once established, should be presumed to continue and that the accused should automatically be confined for treatment until it can be shown that he has recovered”). We cannot say that it was unreasonable and therefore unconstitutional for Congress to make this determination.

The fact that a person has been found, beyond a reasonable doubt, to have committed a criminal act certainly indicates dangerousness.12 See Lynch v. Overholser, 369 U. S. 705, 714 (1962) (The fact that the accused was found to have committed a criminal act is “strong evidence that his continued liberty could imperil ‘the preservation of public peace’ ”). Indeed, this concrete evidence generally may be at least as persuasive as any predictions about dangerousness that might be made in a civil-commitment proceeding.13 We do not agree *365with petitioner’s suggestion that the requisite dangerousness is not established by proof that a person committed a nonviolent crime against property. This Court never has held that “violence,” however that term might be defined, is a prerequisite for a constitutional commitment.14

*366Nor can we say that it was unreasonable for Congress to determine that the insanity acquittal supports an inference of continuing mental illness. It comports with common sense to conclude that someone whose mental illness was sufficient to lead him to commit a criminal act is likely to remain ill and in need of treatment. The precise evidentiary force of the insanity acquittal, of course, may vary from case to case, but the Due Process Clause does not require Congress to make classifications that fit every individual with the same degree of relevance. See Marshall v. United States, 414 U. S. 417, 428 (1974). Because a hearing is provided within 50 days of the commitment, there is assurance that every acquittee has prompt opportunity to obtain release if he has recovered.

Petitioner also argues that, whatever the evidentiary value of the insanity acquittal, the Government lacks a legitimate reason for committing insanity acquittees automatically because it can introduce the insanity acquittal as evidence in a subsequent civil proceeding. This argument fails to consider the Government’s strong interest in avoiding the need to conduct a de novo commitment hearing following every insanity acquittal — a hearing at which a jury trial may be demanded, § 21-544, and at which the Government bears the burden of proof by clear and convincing evidence. Instead of focusing on the critical question whether the acquittee has recovered, the new proceeding likely would have to relitigate much of the criminal trial. These problems accent the Government’s important interest in automatic commitment. See Mathews v. Eldridge, 424 U. S. 319, 348 (1976). We therefore conclude that a finding of not guilty by reason of insanity is a sufficient foundation for commitment of an insanity acquittee for the purposes of treatment and the protection of society.

B

Petitioner next contends that his indefinite commitment is unconstitutional because the proof of his insanity was based only on a preponderance of the evidence, as compared to *367Addington’s civil-commitment requirement of proof by clear and convincing evidence. In equating these situations, petitioner ignores important differences between the class of potential civil-commitment candidates and the class of insanity acquittees that justify differing standards of proof. The Addington Court expressed particular concern that members of the public could be confined on the basis of “some abnormal behavior which might be perceived by some as symptomatic of a mental or emotional disorder, but which is in fact within a range of conduct that is generally acceptable.” 441 U. S., at 426-427. See also O’Connor v. Donaldson, 422 U. S., at 575. In view of this concern, the Court deemed it inappropriate to ask the individual “to share equally with society the risk of error.” Addington, 441 U. S., at 427. But since automatic commitment under § 24-301(d)(1) follows only if the acquittee himself advances insanity as a defense and proves that his criminal act was a product of his mental illness,15 there is good reason for diminished concern as to the risk of error.16 More important, the proof that he committed a criminal act as a result of mental illness eliminates the risk that he is being committed for mere “idiosyncratic behavior,” Addington, 441 U. S., at 427. A criminal act by definition is not “within a range of conduct that is generally acceptable.” Id., at 426-427.

We therefore conclude that concerns critical to our decision in Addington are diminished or absent in the case of insanity acquittees. Accordingly, there is no reason for adopting the same standard of proof in both cases. “[D]ue process is flexible and calls for such procedural protections as the par*368ticular situation demands.” Morrissey v. Brewer, 408 U. S. 471, 481 (1972). The preponderance of the evidence standard comports with due process for commitment of insanity acquittees.17

C

The remaining question is whether petitioner nonetheless is entitled to his release because he has been hospitalized for a period longer than he could have been incarcerated if convicted. The Due Process Clause “requires that the nature and duration of commitment bear some reasonable relation to the purpose for which the individual is committed. ” Jackson v. Indiana, 406 U. S. 715, 738 (1972). The purpose of commitment following an insanity acquittal, like that of civil commitment, is to treat the individual’s mental illness and protect him and society from his potential dangerousness. The committed acquittee is entitled to release when he has recovered his sanity or is no longer dangerous. See O’Connor v. Donaldson, supra, at 575-576; 432 A. 2d, at 372, and n. 16; H. R. Rep. No. 91-907, pp. 73-74 (1970). And because it is impossible to predict how long it will take for any given individual to recover — or indeed whether he ever will recover — Congress has chosen, as it has with respect to civil commitment, to leave the length of commitment indeterminate, subject to periodic review of the patient’s suitability for release.

In light of the congressional purposes underlying commitment of insanity acquittees, we think petitioner clearly errs in contending that an acquittee’s hypothetical maximum sentence provides the constitutional limit for his commitment. A particular sentence of incarceration is chosen to reflect society’s view of the proper response to commission of a par*369ticular criminal offense, based on a variety of considerations such as retribution, deterrence, and rehabilitation. See, e. g., Gregg v. Georgia, 428 U. S. 153, 183-186 (1976) (opinion of Stewart, Powell, and Stevens, JJ.); Kennedy v. Mendoza-Martinez, 372 U. S. 144, 168 (1963); Williams v. New York, 337 U. S. 241, 248-249 (1949). The State may punish a person convicted of a crime even if satisfied that he is unlikely to commit further crimes.

Different considerations underlie commitment of an insanity acquittee. As he was not convicted, he may not be punished.18 His confinement rests on his continuing illness and dangerousness. Thus, under the District of Columbia statute, no matter how serious the act committed by the acquit-tee, he may be released within 50 days of his acquittal if he has recovered. In contrast, one who committed a less serious act may be confined for a longer period if he remains ill and dangerous. There simply is no necessary correlation between severity of the offense and length of time necessary for recovery. The length of the acquittee’s hypothetical criminal sentence therefore is irrelevant to the purposes of his commitment.19

*370<3

We hold that when a criminal defendant establishes by a preponderance of the evidence that he is not guilty of a crime by reason of insanity, the Constitution permits the Government, on the basis of the insanity judgment, to confine him to a mental institution until such time as he has regained his sanity or is no longer a danger to himself or society. This holding accords with the widely and reasonably held view that insanity acquittees constitute a special class that should be treated differently from other candidates for commitment.20 We have observed before that “[w]hen Congress undertakes to act in areas fraught with medical and scientific uncertainties, legislative options must be especially broad and courts should be cautious not to rewrite legislation . . . .” Marshall v. United States, 414 U. S., at 427. This admonition has particular force in the context of legislative efforts to deal with the special problems raised by the insanity defense.

The judgment of the District of Columbia Court of Appeals is

Affirmed.

The Court begins by posing the wrong question. The issue in this case is not whether petitioner must be released because he has been hospitalized for longer than the prison sentence he might have served had he been convicted, any more than the question in a motion to suppress an allegedly coerced confession at a murder trial is whether the murderer should go free.1 The question before us is whether the fact that an individual has been found “not guilty by reason of insanity,” by itself, provides a constitutionally adequate basis for involuntary, indefinite commitment to psychiatric hospitalization.

None of our precedents directly addresses the meaning of due process in the context of involuntary commitments of persons who have been acquitted by reason of insanity. Petitioner’s argument rests primarily on two cases dealing with civil commitments: O’Connor v. Donaldson, 422 U. S. 563 (1975), and Addington v. Texas, 441 U. S. 418 (1979). O’Connor held that a mentally ill individual has a “right to liberty” that a State may not abridge by confining him to a mental institution, even for the purpose of treating his illness, unless in addition to being mentally ill he is likely to harm himself or others if released. 422 U. S., at 573-576; see id., at 589 (Burger, C. J., concurring). Then, in Addington, we carefully evaluated the standard of proof in civil commitment proceedings. Applying the due process analysis of Mathews v. Eldridge, 424 U. S. 319, 335 (1976), *372we held that “due process requires the state to justify confinement by proof more substantial than a mere preponderance of the evidence,” 441 U. S., at 427, specifically “clear and convincing evidence,” id., at 433.2

The core of both cases is a balance of three factors: the governmental interest in isolating and treating those who may be mentally ill and dangerous; the difficulty of proving or disproving mental illness and dangerousness in court; and the massive intrusion on individual liberty that involuntary psychiatric hospitalization entails. Petitioner contends that the same balance must be struck in this case, and that the Government has no greater interest in committing him indefinitely than it has in ordinary civil commitment cases governed by the standards of O’Connor and Addington. While conceding that the Government may have legitimate reasons to commit insanity acquittees for some definite period without carrying the burden of proof prescribed in Addington,3 *373he argues that he cannot be confined indefinitely unless the Government accords him the minimum due process protections required for civil commitment.

A

The obvious difference between insanity acquittees and other candidates for civil commitment is that, at least in the District of Columbia, an acquittal by reason of insanity implies a determination beyond a reasonable doubt that the defendant in fact committed the criminal act with which he was charged. See Bethea v. United States, 365 A. 2d 64, 93-95 (D. C. 1976); D. C. Code §24-301(c) (1981). Conceivably, the Government may have an interest in confining insanity acquittees to punish them for their criminal acts, but the Government disclaims any such interest, and the Court does not rely on it.4 In any event, we have held that the Govern*374ment may not impose psychiatric commitment as an alternative to penal sentencing for longer than the maximum period of incarceration the legislature has authorized as punishment for the crime committed. Humphrey v. Cady, 405 U. S. 504, 510-511 (1972). Once Congress has defined a crime and the punishment for that crime, additional confinement can be justified only by proof beyond a reasonable doubt of additional facts, subject to the limits of the Double Jeopardy Clause, and upon notice to defendants that they are subject to such additional punishment. See Specht v. Patterson, 386 U. S. 605, 610 (1967); In re Winship, 397 U. S. 358, 361-364 (1970).

B

Instead of relying on a punishment rationale, the Court holds that a finding of insanity at a criminal trial “is sufficiently probative of mental illness and dangerousness to justify commitment.” Ante, at 363. First, it declares that “[t]he fact that a person has been found, beyond a reasonable doubt, to have committed a criminal act certainly indicates dangerousness.” Ante, at 364. Second, the Court decides that “[ijt comports with common sense to conclude that someone whose mental illness was sufficient to lead him to commit a criminal act is likely to remain ill and in need of treatment.” Ante, at 366. Despite their superficial appeal, these propositions cannot support the decision necessary to the Court’s disposition of this case — that the Government may be excused from carrying the Addington burden of proof with respect to each of the O’Connor elements of mental illness and dangerousness in committing petitioner for an indefinite period.

*3751. Our precedents in other commitment contexts are inconsistent with the argument that the mere facts of past criminal behavior and mental illness justify indefinite commitment without the benefits of the minimum due process standards associated with civil commitment, most importantly proof of present mental illness and dangerousness by clear and convincing evidence.5 In Addington itself, the petitioner did not dispute that he had engaged in a wide variety of assaultive conduct that could have been the basis for criminal charges had the State chosen to prosecute him. See 441 U. S., at 420-421. Similarly, the petitioner in Jackson v. Indiana, 406 U. S. 715 (1972), had been charged with two robberies, yet we required the State to follow its civil commitment procedures if it wished to commit him for more than a strictly limited period. Id., at 729-730. As the Court indicates, see ante, at 364, n. 12, these cases are perhaps distinguishable on the ground that there was never proof that a crime had been committed, although in Addington the petitioner’s violent acts were before the jury. That objection, however, cannot be leveled at Baxstrom v. Herold, 383 U. S. 107 (1966), or Humphrey v. Cady, supra.

The petitioner in Baxstrom had been convicted of assault and sentenced to a term in prison, during which he was certified as insane by a prison physician. At the expiration of his criminal sentence, he was committed involuntarily to a state mental hospital under procedures substantially less protective than those used for civil commitment. 383 U. S., at *376108-110. We held that, once he had served his sentence, Baxstrom could not be treated differently from other candidates for civil commitment. Id., at 112-113. The principal difference between this case and Baxstrom is petitioner’s admission, intrinsic to an insanity plea in the District of Columbia at the time of his trial, that his crime was “the product” of his mental illness. Humphrey, however, indicates the limited importance of that distinction.

In Humphrey, the petitioner had been convicted of contributing to the delinquency of a minor, the court had determined that his crime was “probably directly motivated by a desire for sexual excitement,” and the State had established his “need” for psychiatric treatment by a preponderance of the evidence at a special hearing. 405 U. S., at 506-507. He was committed for treatment for the maximum period for which he could have been incarcerated as punishment for his crime — as in this case, one year — and at the end of that period his commitment was renewed for five more years after a judicial hearing on his present mental illness and dangerousness. See id., at 507. Thus, the situation was almost precisely identical to that in this case after petitioner’s February 1977 hearing — the defendant had been found to have committed a criminal act beyond a reasonable doubt, a connection between that act and a mental disorder had been established by a preponderance of the evidence, and he had been confined for longer than the maximum sentence he could have received. If anything, Humphrey had received more protections than Michael Jones; the State had borne the burden of proof by a preponderance of the evidence at his “release hearing,” ibid., and his recommitment was for a strictly limited time. Nevertheless, we held that Humphrey’s constitutional challenge to the renewal order had substantial merit, because Humphrey had not received the procedural protections given persons subject to civil commitment.6

*3772. The Government’s interests in committing petitioner are the same interests involved in Addington, O’Connor, Baxstrom, and Humphrey — isolation, protection, and treatment of a person who may, through no fault of his own, cause harm to others or to himself. Whenever involuntary commitment is a possibility, the Government has a strong interest in accurate, efficient commitment decisions. Nevertheless, Addington held both that the government’s interest in accuracy was not impaired by a requirement that it bear the burden of persuasion by clear and convincing evidence, and that the individual’s interests in liberty and autonomy required the government to bear at least that burden. An acquittal by reason of insanity of a single, nonviolent misdemeanor is not a constitutionally adequate substitute for the due process protections of Addington and O’Connor, i. e., proof by clear and convincing evidence of present mental illness or dangerousness, with the government bearing the burden of persuasion.

A “not guilty by reason of insanity” verdict is backward-looking, focusing on one moment in the past, while commitment requires a judgment as to the present and future. In some jurisdictions, most notably in federal criminal trials, an acquittal by reason of insanity may mean only that a jury found a reasonable doubt as to a defendant’s sanity and as to the causal relationship between his mental condition and his crime. See Davis v. United States, 160 U. S. 469 (1895). As we recognized in Addington, “[t]he subtleties and nuances *378of psychiatric diagnosis render certainties virtually beyond reach in most situations.” 441 U. S., at 430. The question is not whether “government may not act in the face of this uncertainty,” ante, at 365, n. 13; everyone would agree that it can. Rather, the question is whether — in light of the uncertainty about the relationship between petitioner’s crime, his present dangerousness, and his present mental condition— the Government can force him for the rest of his life “to share equally with society the risk of error,” 441 U. S., at 427.7

It is worth examining what is known about the possibility of predicting dangerousness from any set of facts. Although a substantial body of research suggests that a consistent pattern of violent behavior may, from a purely statistical standpoint, indicate a certain likelihood of further violence in the future,8 mere statistical validity is far from perfect for purposes of predicting which individuals will be dangerous. Commentators and researchers have long acknowledged that even the best attempts to identify dangerous individuals on the basis of specified facts have been inaccurate roughly two-thirds of the time, almost always on the side of overprediction.9 On a clinical basis, mental health profes*379sionals can diagnose past or present mental condition with some confidence, but strong institutional biases lead them to err when they attempt to determine an individual’s dangerousness, especially when the consequence of a finding of dangerousness is that an obviously mentally ill patient will remain within their control.10 Research is practically nonexistent on the relationship of nonviolent criminal behavior, such as petitioner’s attempt to shoplift, to future dangerousness. We do not even know whether it is even statistically valid as a predictor of similar nonviolent behavior, much less of behavior posing more serious risks to self and others.

Even if an insanity acquittee remains mentally ill, so long as he has not repeated the same act since his offense the passage of time diminishes the likelihood that he will repeat it.11 Furthermore, the frequency of prior violent behavior is an important element in any attempt to predict future violence.12 Finally, it cannot be gainsaid that some crimes are more indicative of dangerousness than others. Subject to the limits of O’Connor, a State may consider nonviolent misdemeanors “dangerous,” but there is room for doubt whether a single attempt to shoplift and a string of brutal murders are equally *380accurate and equally permanent predictors of dangerousness.13 As for mental illness, certainly some conditions that satisfy the “mental disease” element of the insanity defense do not persist for an extended period — thus the traditional inclusion of “temporary insanity” within the insanity defense.

Close reading of the Court’s opinion reveals the utter emptiness of the legislative judgment it finds so unproblematic. Today’s decision may overrule Humphrey by implication. It does not, however, purport to overrule Baxstrom or any of the cases which have followed Baxstrom.14 It is clear, therefore, that the separate facts of criminality and mental illness cannot support indefinite psychiatric commitment, for both were present in Baxstrom. The Court’s careful phrasing indicates as much: “someone whose mental illness was sufficient to lead him to commit a criminal act is likely to remain ill and in need of treatment.” Ante, at 366 (emphasis added). The Court relies on a connection between mental condition and criminal conduct that is unique to verdicts of “not guilty by reason of insanity.” Yet the relevance of that connection, as opposed to each of its separate components, is far from a matter of obvious “common sense.” None of the available evidence that criminal behavior by the mentally ill is likely to repeat itself distinguishes between behaviors that were “the product” of mental illness and those that were not.15 It is *381completely unlikely that persons acquitted by reason of insanity display a rate of future “dangerous” activity higher than civil committees with similar arrest records, or than persons convicted of crimes who are later found to be mentally ill. The causal connection between mental condition and criminal behavior that “not guilty by reason of insanity” formulations universally include is more a social judgment than a sound basis for determining dangerousness.

Given the close similarity of the governmental interests at issue in this case and those at issue in Addington, and the highly imperfect “fit” between the findings required for an insanity acquittal and those required under O’Connor to support an indefinite commitment, I cannot agree that the Government should be excused from the burden that Addington held was required by due process.16

3. In considering the requirements of due process, we have often inquired whether alternative procedures more protective of individual interests, at a reasonable cost, were likely to accomplish the State’s legitimate objectives. See, *382e. g., Mathews v. Eldridge, 424 U. S., at 335; Stanley v. Illinois, 405 U. S. 645, 657-658 (1972); Bell v. Burson, 402 U. S. 535, 542-543 (1971). There are many ways to take into account criminal behavior and past mental condition, and thereby to vindicate the government’s legitimate interest in accurate commitment decisions, without depriving insanity acquittees of the Addington protections. Certain aspects of the District of Columbia’s commitment procedures already embody less restrictive alternatives: all insanity acquittees are committed automatically for 50 days before an initial release hearing, § 24-301(d), and the testimony of mental health professionals at all hearings may be informed by their experience with mentally ill patients and by their familiarity with current research. The fact of an insanity acquittal and the evidence on insanity adduced at trial are clearly admissible in all commitment and release hearings.

In addition, an insanity acquittal might conceivably justify commitment for a reasonably limited period without requiring the Government to meet its Addington burden. See United States v. Brown, 155 U. S. App. D. C. 402, 408, 478 F. 2d 606, 612 (1973); American Psychiatric Assn., Statement on the Insanity Defense 15 (1982); cf. Jackson v. Indiana, 406 U. S., at 738; McNeil v. Director, Patuxent Institution, 407 U. S. 245, 249 (1972). In this case, petitioner submits that such a reasonable period extends no longer than the maximum sentence that could have been imposed had he been found guilty of the crime charged. But at some point the Government must be required to justify further commitment under the standards of Addington. 17

*3834. If the Government’s interests were the only ones at stake, an insanity acquittal would furnish a reasonable basis for indefinite commitment. Under the Constitution, however, the Government’s interests must be considered in light of the liberty interests of the individual who is subject to commitment. In the final analysis, the Court disregards Addington not on the ground that the Government’s interests in committing insanity acquittees are different from or stronger than its interests in committing criminals who happen to be mentally ill, or mentally ill individuals who have done violent, dangerous things, but on the theory that “there is good reason for diminished concern as to the risk of error” when a person is committed indefinitely on the basis of an insanity acquittal. See ante, at 367.

The “risk of error” that, according to the Court, is diminished in this context subsumes two separate risks. First, the Court notes that in Addington we were concerned, at least in part, that individuals might be committed for mere idiosyncratic behavior, see 441 U. S., at 427, and it observes that criminal acts are outside the “ ‘range of conduct that is generally acceptable.’” Ante, at 367, quoting 441 U. S., at 426-427. O’Connor, however, requires that a person be proved dangerous, not merely “unacceptable,” before he may *384be subjected to the massive curtailment of individual freedom and autonomy that indefinite commitment entails. In Addington itself, the State had clearly proved by a preponderance of the evidence that the petitioner had engaged repeatedly in conduct far beyond the pale of acceptable behavior, yet we did not regard that level of proof as furnishing adequate protection for the individual interests at stake.18

Second, the Court reasons that “[a] criminal defendant who successfully raises the insanity defense necessarily is stigmatized by the verdict itself,” and therefore that committing him does not involve the same risk of stigmatization a civil commitment may entail. Ante, at 367, n. 16. This is perhaps the Court’s most cynical argument. It is true that in Addington and in Vitek v. Jones, 445 U. S. 480 (1980), we recognized that individuals have an interest in not being stigmatized by society at large on account of being labeled mentally ill. 441 U. S., at 426; 445 U. S., at 492. Avoiding stigma, however, is only one of the reasons for recognizing a liberty interest in avoiding involuntary commitment. We have repeatedly acknowledged that persons who have already been labeled as mentally ill nonetheless retain an interest in avoiding involuntary commitment. See, e. g., O’Connor v. Donaldson, 422 U. S., at 575; Baxstrom v. Herold, 383 U. S. 107 (1966). Other aspects of involuntary commitment affect them in far more immediate ways.

In many respects, confinement in a mental institution is even more intrusive than incarceration in a prison. Inmates of mental institutions, like prisoners, are deprived of unrestricted association with friends, family, and community; *385they must contend with locks, guards, and detailed regulation of their daily activities. In addition, a person who has been hospitalized involuntarily may to a significant extent lose the right enjoyed by others to withhold consent to medical treatment. See Youngberg v. Romeo, 457 U. S. 307, 321 (1982) (involuntary committee’s due process right to freedom from unreasonable restraint limited to a guarantee that professional medical judgment be exercised). The treatments to which he may be subjected include physical restraints such as straightjacketing, as well as electroshock therapy, aversive conditioning, and even in some eases psychosurgery. Administration of psychotropic medication to control behavior is common. See American Psychiatric Assn., Statement on the Insanity Defense 15 (1982) (“Greater emphasis is now placed upon psychopharmacological management of the hospitalized person”). Although this Court has never approved the practice, it is possible that an inmate will be given medication for reasons that have more to do with the needs of the institution than with individualized therapy.19 See Mills v. Rogers, 457 U. S. 291, 303 (1982); Rennie v. Klein, 653 F. 2d 836, 845 (CA3 1981) (en banc). We should not presume that he lacks a compelling interest in having the decisions to com*386mit him and to keep him institutionalized made carefully, and in a manner that preserves the maximum degree of personal autonomy.

Therefore, I cannot agree with the Court that petitioner in this case has any less interest in procedural protections during the commitment process than the petitioners in Addington, O’Connor, or Baxstrom, and I cannot agree that the risks of error which an indefinite commitment following an insanity acquittal entails are sufficiently diminished to justify relieving the Government of the responsibilities defined in Addington.

C

Indefinite commitment without the due process protections adopted in Addington and O’Connor is not reasonably related to any of the Government’s purported interests in confining insanity acquittees for psychiatric treatment. The rationales on which the Court justifies § 24-301’s departures from Addington at most support deferring Addington’s, due process protections — specifically, its requirement that the Government carry the burden of proof by clear and convincing evidence — for a limited period only, not indefinitely.

The maximum sentence for attempted petit larceny in the District of Columbia is one year. Beyond that period, petitioner should not have been kept in involuntary confinement unless he had been committed under the standards of Addington and O’Connor. Petitioner had been in custody for 17 months at the time of his February 1977 hearing, either in St. Elizabeths or in the District of Columbia Correctional Center. At that time he should have received the benefit of the Addington due process standards, and, because he did not, the findings at that hearing cannot provide constitutionally adequate support for his present commitment. I would therefore reverse the judgment of the District of Columbia Court of Appeals.

dissenting.

The character of the conduct that causes a person to be incarcerated in an institution is relevant to the length of his permissible detention. In my opinion, a plea of not guilty by reason of insanity, like a plea of guilty, may provide a sufficient basis for confinement for the period fixed by the legislature as punishment for the acknowledged conduct, provided of course that the acquittee is given a fair opportunity to prove that he has recovered from his illness. But surely if he is to be confined for a longer period, the State must shoulder the burden of proving by clear and convincing evidence that such additional confinement is appropriate. As Justice Brennan demonstrates, that result is dictated by our prior cases. What Justice Powell has written lends support to the view that the initial confinement of the acquittee is permissible, but provides no support for the conclusion that he has the burden of proving his entitlement to freedom after he has served the maximum sentence authorized by law. I respectfully dissent because I believe this shoplifter was presumptively entitled to his freedom after he had been incarcerated for a period of one year.