Opinion
KENNARD, J.Resolving a question we left open in People v. Torres (2001) 25 Cal.4th 680, 686, footnote 2 [106 Cal.Rptr.2d 824, 22 P.3d 871], *1182we hold that before a defendant can be committed or recommitted under the Sexually Violent Predators Act (Welf. & Inst. Code, § 6600 et seq. (SVPA)),1 the trier of fact must find, beyond a reasonable doubt, that the defendant is likely to commit sexually violent predatory behavior upon release. Thus, the judge or jury trying the case must determine not only whether the defendant is likely to “engage in sexually violent criminal behavior” (§ 6600, subd. (a)), but also whether that behavior is likely to be directed “toward a stranger, a person of casual acquaintance with whom no substantial relationship exists, or an individual with whom a relationship has been established or promoted for the primary purpose of victimization.” (§ 6600, subd. (e).)
The Court of Appeal here reached the same conclusion, and held that the trial court erred in failing to instruct the jury to determine whether defendant was likely to commit future predatory acts. The Court of Appeal also concluded, however, that under the facts of this case the error was harmless beyond a reasonable doubt. We agree with the Court of Appeal opinion in all respects, and affirm its judgment.
I
As in People v. Torres, supra, 25 Cal.4th 680 {Torres), the issue raised here stems from the SVPA’s use of dissimilar language to define a “predatory act” and a “sexually violent predator.” We observed in Torres: “A predatory act is one directed ‘toward a stranger, a person of casual acquaintance with whom no substantial relationship exists, or an individual with whom a relationship has been established or promoted for the primary purpose of victimization.’ (§ 6600, subd. (e).) In contrast, the [SVPA] does not define a sexually violent predator by reference to victim characteristics. Rather, a sexually violent predator is ‘a person who has been convicted of a sexually violent offense against two or more victims and who has a diagnosed mental disorder that makes the person a danger to the health and safety of others in that it is likely that he or she will engage in sexually violent criminal behavior.’ (§ 6600, subd. (a).) In turn, a ‘[s]exually violent offense’ under the [SVPA] includes forcible rape, forcible spousal rape, rape in concert, lewd act against a person under 14 years of age, forcible penetration, forcible sodomy, and forcible oral copulation. (§ 6600, subd. (b)-)
“Under section 6601, whenever the Director of Corrections determines that a defendant serving a prison term may be a sexually violent predator, the Department of Corrections and the Board of Prison Terms undertake an *1183initial screening ‘based on whether the person has committed a sexually violent predatory offense and on a review of the person’s social, criminal, and institutional history.’ (§ 6601, subd. (b).) The screening is conducted in accord with a protocol developed by the state Department of Mental Health. (Ibid.) If that screening leads to a determination that the defendant is likely to be a sexually violent predator, the defendant is referred to the Department of Mental Health for an evaluation by two psychiatrists or psychologists. (§ 6601, subds. (b) & (c).) If both find that the defendant ‘has a diagnosed mental disorder so that he or she is likely to engage in acts of sexual violence without appropriate treatment and custody’ (§ 6601, subd. (d)), the department forwards a petition for commitment to the county of the defendant’s last conviction (ibid.). If the county’s designated counsel concurs with the recommendation, he or she files a petition for commitment in the superior court. (§ 6601, subd. (i).)” (Torres, supra, 25 Cal.4th at pp. 682-683.)
The statute’s definition of “predatory act” comes into play at the next stage, the probable cause hearing. There, the superior court must “determine whether there is probable cause to believe that the [defendant] is likely to engage in sexually violent predatory criminal behavior upon his or her release.” (§ 6602, subd. (a), italics added.) If the court finds probable cause, the matter is set for trial.
Section 6604 provides that at trial “[t]he court or jury shall determine whether, beyond a reasonable doubt, the person is a sexually violent predator.” The phrase “sexually violent predator” is defined in section 6600, subdivision (a), as a person likely to “engage in sexually violent criminal behavior.” Nothing in section 6604 expressly requires the trier of fact to determine whether the sexually violent criminal behavior is “predatory” behavior—that is, behavior directed at a stranger, a casual acquaintance, or someone cultivated for victimization as defined in section 6600, subdivision (e). Thus, as we noted in Torres, the SVPA is “unusual in that its language requires the court at a probable cause hearing to decide whether the defendant ‘is likely to engage in sexually violent predatory criminal behavior upon his or her release’ (§ 6602, subd. (a)), but it does not expressly provide for such an issue to be decided by the trier of fact at the trial.” (Torres, supra, 25 Cal.4th at p. 686.)
II
We take the facts from the Court of Appeal opinion:
“In 1987, [defendant] Hurtado sodomized and committed lascivious acts on three minor boys (ages 14, 13 and 12). (Pen. Code, §§ 288, subd. (a), 286, *1184subd. (c).) He approached the minors at a Pizza Hut while they were playing video games, took them to his residence where he and the victims engaged in various sex acts and threatened to kill them if they told anyone. He was sentenced to 17 years in prison with a scheduled parole release date of August 26, 1996.
“[Defendant] had earlier prior convictions for sexual misconduct. In 1979, he was convicted of lewd or lascivious conduct upon a child under the age of 14. He was adjudicated as a mentally disordered sex offender and sent to Atascadero State Hospital. After one year, he was deemed unamenable to treatment and returned to the trial court for the resumption of criminal proceedings. He received a sentence of seven years and was paroled on June 1, 1984. [Defendant] violated that parole and returned to prison for another year. In 1985, he again violated parole and was convicted of annoying and molesting children. He was sentenced to prison for another two years. Less than a week after being released, [defendant] violated his parole when he was found in the presence of a minor without adult supervision. He was sent back to prison until August 1987. Less than two months later, he was arrested for the offenses with three minor boys summarized above.
“After [defendant] was evaluated by two clinical psychologists who concurred he met the [SVPA]’s criteria for commitment as a sexually violent predator, the Department of Mental Health on August 15, 1996, referred [defendant’s case to the district attorney for filing of a civil commitment petition pursuant to section 6602. A commitment petition was filed on October 9. That petition was dismissed without prejudice and a new one was filed on October 29, alleging [defendant] was a sexually violent predator who was likely to engage in sexually violent criminal behavior upon his release as a result of his diagnosed mental disorder. A probable cause hearing began on December 18, 1996, and a finding of probable cause was returned in January 3, 1997.
“A jury trial began on September 16. Dr. Harry Goldberg, a clinical psychologist who had evaluated [defendant] in July 1996, testified [defendant] suffered from two diagnosed mental disorders, to wit, pedophilia and antisocial personality. During the interview, [defendant] had admitted participating in the sexual acts with the children and fantasizing about children. Dr. Goldberg testified that his testing placed [defendant] within the severe range of psychopathy, explaining that a high score is predictive of violence and poor parole adjustment in the future. He also stated that because those inmates who focus on males rather than females have a higher recidivism rate, [defendant’s preference for sexual offenses with males indicated a higher risk of reoffending. He believed [defendant] was predisposed to engage in sexually violent behavior in the future.
*1185“Dr. Dean Clair, a clinical psychologist who evaluated [defendant] in June 1996, concluded the 1987 offenses fit the criteria of section 6600, in that they were predatory and violent offenses involving children whom [defendant] sought out and who were unknown to him. Dr. Clair concluded [defendant] had a diagnosed mental disorder of pedophilia, primarily basing that determination on [defendant]’s history of aggressive sexuality toward victims 13 years of age and younger and recidivism. Dr. Clair concluded that based on the length and history of [defendant]’s condition, his behavioral recidivism, the nature of pedophilia and his inability to perform on parole, he was at considerable risk of reoffending.”
The trial court instructed the jury under CALJIC No. 4.19, as it read before the 2002 revision.2 The instruction directed the jury to determine whether defendant was a “sexually violent predator,” and it defined that term as referring to someone who is “a danger to the health and safety of others in that it is likely that he or she will engage in sexually violent criminal behavior.” The instruction included the statutory definition of predatory act as an act “directed towards a stranger, a person of casual acquaintance with whom no substantial relationship exists, or individual with whom a relationship has been established or promoted for the primary purpose of victimization.” (§ 6600, subd. (e).) But that definition served no function, because the jury was not asked to determine whether defendant was likely to commit predatory acts.
The jury found that defendant was a sexually violent predator as defined in section 6600. He was committed to the state Department of Mental Health for placement in a secure facility for a period not to exceed two years.
Defendant appealed. While his appeal was pending before the Court of Appeal, his two-year commitment expired, and he was recommitted in a new proceeding. The Court of Appeal nevertheless retained the appeal, and held that the trial court erred in failing to instruct the jury that it could not return a verdict that defendant was a sexually violent predator unless it found that he was likely to engage in predatory behavior. The Court of Appeal further concluded, however, that the error was harmless beyond a reasonable doubt.
Defendant and the Attorney General both petitioned for review. We granted both petitions, and we held this case pending our decision in Torres, supra, 25 Cal.4th 680. We decided Torres without resolving the conflict in *1186the Court of Appeal decisions as to whether a jury in an SVPA case must determine whether the defendant is likely to commit future predatory acts. We therefore set this case for briefing and argument to address that issue.
Although this case became moot when defendant’s 1996 commitment expired in 1998, we have retained it for decision because the issue it raises is recurrent—indeed, it is raised in virtually every SVPA trial and appeal— and the two-year limit on each commitment makes it likely that any appeal raising the issue would become moot before we could decide it. (See People v. Cheek (2001) 25 Cal.4th 894, 897-898 [108 Cal.Rptr.2d 181, 24 P.3d 1204]; Conservatorship of Wendland (2001) 26 Cal.4th 519, 524, fn. 1 [110 Cal.Rptr.2d 412, 28 P.3d 151].)
Ill
We conclude that section 6604 contains an implied requirement that a trier of fact must find beyond a reasonable doubt that the defendant is likely to commit sexually violent predatory criminal acts before the defendant can be committed as a sexually violent predator. We base this conclusion on the structure of the SVPA and the purpose of the April 17, 1995, amendments that added “predatory act” language to the bill that became the SVPA.
The SVPA provides for both a preliminary probable cause hearing and a later trial. In In re Parker (1998) 60 Cal.App.4th 1453, 1468 [71 Cal.Rptr.2d 167], the Court of Appeal noted that the Legislature modeled the probable cause hearing of section 6602 after existing involuntary civil commitment procedures and criminal preliminary hearing procedures. As the Court of Appeal below stated: “Regardless of the legal context of a probable cause determination (an involuntary civil commitment or a criminal preliminary hearing), its role is to test the sufficiency of the factual basis underlying the People’s pleadings.” It would be inconsistent with the purpose of a probable cause hearing to require the judge to determine whether there was probable cause to believe the People can prove a factual assertion that they are not required to prove at trial. Thus the SVPA’s provision for a preliminary hearing to test whether there is probable cause to believe a defendant is likely to commit future predatory acts implies that the likelihood of such acts will also be an issue at trial.
This conclusion is supported by the purpose of the SVPA as revealed through its legislative history. The bill that ultimately became law as the SVPA was introduced in the California Assembly on February 22, 1995, as Assembly Bill No. 888 (1995-1996 Reg. Sess.). That bill initially defined a *1187sexually violent predator as one who engages in sexually violent criminal behavior, and neither defined nor used the term “predatory acts.”
On April 17, 1995, Assembly Bill No. 888 (1995-1996 Reg. Sess.) was amended to add a definition of the term “predatory” to section 6600. The amendment inserted the phrase “sexually violent predatory criminal behavior” in section 6602, which describes the probable cause hearing. The amendment also revised section 6601, subdivision (b) to provide that the initial screening to determine if SVPA proceedings were warranted should include review of whether the defendant has committed a “sexually violent predatory offense,” and section 6607, subdivision (a), to provide that the Director of Mental Health may recommend conditional release of a person not likely to commit acts of predatory sexual violence.3
The apparent purpose—indeed, the only purpose we can discern—of these amendments is to limit the class of persons subject to commitment to those who are likely to commit predatory acts. The analysis of Assembly Bill No. 888 (1995-1996 Reg. Sess.), after its amendment on April 17, 1995, stated that the amended bill would define “a sexually violent predator as a person who has been convicted of a sexually violent offense and who has a mental abnormality or personality disorder that makes the person a danger to the health and safety of others in that it is likely that he or she will engage in sexually violent predatory criminal behavior.” (Assem. Com. on Public Safety, Rep. on Assem. Bill No. 888 (1995-1996 Reg. Sess.) as amended Apr. 17, 1995, italics added.) The same assessment appears in a subsequent analysis. (Assem. Com. on Public Safety, 3d reading analysis of Assem. Bill No. 888 (1995-1996 Reg. Sess.) as amended May 31, 1995, p. 1.)
The class of persons who commit predatory acts as defined in section 6600, subdivision (e), is a much smaller group than the class of persons who commit sexually violent criminal acts as defined in the same section. It includes, however, the most dangerous offenders. Because predatory offenders could strike at any time and victimize anyone, they pose a much greater threat to the public at large. In contrast, a defendant likely to commit crimes only against family members or close acquaintances is less likely to reoffend because potential victims will be, aware of the defendant’s status as a sex *1188offender. The public at large, however, is inevitably more defenseless against acts committed by strangers.
The Attorney General agrees that the purpose of the amendment was to limit commitment to those persons likely to commit predatory acts—those directed “toward a stranger, a person of casual acquaintance with whom no substantial relationship exists, or an individual with whom a relationship has been established or promoted for the primary purpose of victimization.” (§ 6600, subd. (e).) He contends, however, that the Legislature intended to assign responsibility for excluding offenders unlikely to commit predatory acts to the judge at the preliminary hearing. Taking the issue from the jury, he argues, would spare jurors the moral anguish of having to reject commitment for a defendant likely to commit violent sexual crimes simply because the defendant’s likely victims were family members or close acquaintances. Moreover, he argues, the judge is better able to evaluate psychiatric testimony and records than a jury.
The Attorney General’s conjecture concerning legislative intent is creative, but unsupported by anything in the legislative history. Jurors are often called upon to make morally troubling decisions; they are required, for example, to acquit a defendant if not convinced of guilt beyond a reasonable doubt even though they believe the defendant probably committed a serious crime and will probably strike again. Jurors are also often required to evaluate psychiatric evidence. Indeed, even under the Attorney General’s proposed reading the jurors will still have to examine psychiatric evidence to determine if the defendant is likely to commit future acts of sexual violence against any class of victim, and may face the moral anguish of having to reject commitment for a defendant likely to reoffend if the state cannot prove its case beyond a reasonable doubt.
The principal difficulty with the Attorney General’s argument, however, is not that under his construction of the SVPA a judge instead of a jury makes the decision whether the defendant is likely to commit future predatory acts, but that the judge makes this decision at a probable cause hearing instead of at the trial.4 The SVPA does not set out a definition of “probable cause,” but “the rule of law is well established that where the legislature uses terms already judicially construed, ‘the presumption is almost irresistible that it used them in the precise and technical sense which had been placed upon them by the courts.’” (City of Long Beach v. Marshall (1938) 11 Cal.2d 609, 620 [82 P.2d 362].) The term “probable cause” has an established *1189meaning in connection with criminal proceedings, and signifies a level of proof below that of proof beyond a reasonable doubt, or even proof by a preponderance of the evidence. It refers to “a state of facts as would lead a man of ordinary caution and prudence to believe and conscientiously entertain a strong suspicion of the guilt of the accused.” (People v. Uhlemann (1973) 9 Cal.3d 662, 667 [108 Cal.Rptr. 657, 511 P.2d 609]; see Cummiskey v. Superior Court (1992) 3 Cal.4th 1018, 1027 [13 Cal.Rptr.2d 551, 839 P.2d 1059]; People v. Fischer (1957) 49 Cal.2d 442, 446 [317 P.2d 967] [probable cause to arrest]; Williams v. Superior Court (1969) 71 Cal.2d 1144, 1147 [80 Cal.Rptr. 747, 458 P.2d 987]; 4 Witkin & Epstein, Cal. Criminal Law (3d ed. 2000) Pretrial Proceeding, § 147, pp. 349-350.) This definition of “probable cause” is so well settled that we can confidently assume the Legislature intended it to be used in proceedings under section 6602, subdivision (a).
Thus, at the SVPA probable cause hearing the court is required to decide whether it has a strong suspicion that the defendant, by reason of a diagnosed mental disorder, is (1) “likely to engage in acts of sexual violence upon his or her release” (§ 6602, subd. (a)); and (2) is likely to engage in “sexually violent predatory criminal behavior upon his or her release” (ibid.) (italics added). A reasonable suspicion standard is appropriate to the first issue, because the effect of a finding of probable cause to believe the defendant is likely to engage in future acts of sexual violence is to submit that issue for trial under a reasonable doubt standard. Under the Attorney General’s proposed construction, however, the second issue—whether the defendant is likely to engage in future predatory acts—does not become an issue for trial but is resolved by a finding of probable cause. Thus a defendant could be committed if there were no more than a reasonable suspicion that the defendant was likely to commit future predatory acts, even if that suspicion was against the preponderance of the evidence.
The Legislative purpose of limiting commitment under the SVPA to only those persons who present a “substantial danger” (People v. Superior Court (Ghilotti) (2002) 27 Cal.4th 888, 922 [119 Cal.Rptr.2d 1, 44 P.3d 949]) of committing future predatory acts would be subverted if defendants could be committed when there is no more than a reasonable suspicion that they would commit such acts. One commentator explained: “[I]f the standard to be used for [commitment] ... is that there is a ‘likelihood the individual will engage in sexually violent criminal behavior,’ then the Act’s narrowing function has essentially been eviscerated. This standard is broader and looser than the one used for a probable cause determination, and necessarily would permit a larger number of individuals to be involuntarily committed than a standard which required a likelihood of the commission of a sexually violent predatory act. It would also nullify the safeguards sewn into the fabric of the *1190Act since the narrowing process . . . would be eliminated by the use of this more general standard, in violation of due process protection.” (Comparet-Cassani, A Primer on the Civil Trial of a Sexually Violent Predator (2000) 37 San Diego L.Rev. 1057, 1107.)
An interpretation of the SVPA that limits the predatory act determination to the probable cause hearing also creates problems with judicial review. In People v. Talhelm (2000) 85 Cal.App.4th 400, 405 [102 Cal.Rptr.2d 150], the Court of Appeal applied to an SVPA case the rule of People v. Pompa-Ortiz (1980) 27 Cal.3d 519, 529 [165 Cal.Rptr. 851, 612 P.2d 941], that errors at a preliminary hearing are not reviewable on appeal unless they deprived the defendant of a fair trial or otherwise caused prejudice. That rule rests on the belief that a fair trial generally renders harmless any preliminary hearing errors. (See People v. Millwee (1998) 18 Cal.4th 96, 122 [74 Cal.Rptr.2d 418, 954 P.2d 990].) Such reasoning is inapplicable, however, if an issue at the preliminary hearing is excluded from consideration at trial.
In sum, to permit a final determination of the predatory act issue at the probable cause hearing under a reasonable suspicion standard would undermine the Legislature’s plan to limit commitment to persons likely to commit such acts. The more reasonable construction of the SVPA, consistent with its structure and purpose, is that if a judge at the section 6602 hearing finds probable cause to believe a defendant is likely to engage in future predatory acts, that issue then becomes one for decision by the trier of fact under the statutory standard of proof beyond a reasonable doubt.
IV
Having decided that the trial court here erred in not instructing the jury on the predatory act requirement, we turn to the question of whether that error was prejudicial. As we have noted, this case became moot once the defendant’s two-year commitment expired. But the standard of prejudice for constitutional error in SVPA cases, like the instructional issue discussed earlier, is an important and recurring issue likely to evade review, one raised in many cases awaiting this decision. We therefore exercise our discretion to decide this issue.
The United States Supreme Court has not spoken on the standard of prejudice for instructional error in involuntary commitment cases, leaving the parties here free to argue alternative theories. Defendant contends the failure to instruct the jury on the predatory act requirement is “structural *1191error” that is not susceptible to harmless error analysis, but requires automatic reversal. (See Sullivan v. Louisiana (1993) 508 U.S. 275 [113 S.Ct. 2078, 124 L.Ed.2d 182] [defective reasonable doubt instruction].) The Attorney General argues for the test of prejudice we articulated in People v. Watson (1956) 46 Cal.2d 818 [299 P.2d 243], which applies to all nonstructural state law error (People v. Cahill (1993) 5 Cal.4th 478, 492 [20 Cal.Rptr.2d 582, 853 P.2d 1037]), and requires a defendant to show that without the error a more favorable outcome was reasonably probable. The Court of Appeal took an intermediate position, applying the standard of Chapman v. California (1967) 386 U.S. 18 [87 S.Ct. 824, 17 L.Ed.2d 705, 24 A.L.R.3d 1065], which applies to federal constitutional error and requires the state to show that the error is harmless beyond a reasonable doubt.
A recent decision of the United States Supreme Court casts a shadow over defendant’s contention that failure to instruct on an element essential to commitment is structural error requiring automatic reversal. In Neder v. United States (1999) 527 U.S. 1 [119 S.Ct. 1827, 144 L.Ed.2d 35], the trial court erroneously denied the defendant the right to have a jury decide whether an alleged fraudulent statement was “material,” and instead decided that issue itself. The high court recognized that the error deprived the defendant of a constitutional right, but after reviewing precedent concluded “the omission of an element is subject to harmless-error analysis” (id. at p. 10 [119 S.Ct. at p. 1834]) and found the error harmless under the Chapman standard (Chapman v. California, supra, 386 U.S. 18).
Defendant argues his case falls within an exception to this rule. His theory is that harmless error analysis requires the court to review the record to evaluate the effect of the error on the result, and that accurate review is impossible if the parties, unaware that an issue was before the court, did not present evidence on that issue. He relies on Lankford v. Idaho (1991) 500 U.S. 110 [111 S.Ct. 1723, 114 L.Ed.2d 173], in which a judge imposed a death penalty even though the prosecution had assured the defense that it was not seeking the death penalty. Under Idaho law in 1991, the jury decided guilt in a capital case and the judge, after a sentencing hearing, determined the penalty. (See Idaho Code, § 19-2515; but see Ring v. Arizona (2002) 536 U.S. 584 [122 S.Ct. 2428, 153 L.Ed.2d 556].) The high court reversed the penalty. It could not rely on the record to hold the error harmless because the defense, trusting the prosecution’s assurance, did not present evidence on aggravating or mitigating factors relating to the penalty decision. (See Lankford v. Idaho, supra, 500 U.S. at pp. 122-127 [111 S.Ct. at pp. 1730-1733].)
The case before us differs from Lanlford. Defendant here knew that the likelihood of future predatory acts was at issue in the probable cause *1192hearing. He had an opportunity to present evidence on predatory acts to a judge. (See In re Parker, supra, 60 Cal.App.4th at pp. 1469-1470 [a defendant has a right to present evidence at probable cause hearing]; People v. Cheek, supra, 25 Cal.4th at p. 900 [endorsing Parker].) Here, the evidence actually presented at defendant’s trial—the nature of his prior acts, his prison record, and the psychiatric reports—is in large part the same evidence that a jury would consider in finding whether defendant was likely to commit future predatory acts. Defendant does not suggest that he has any additional evidence or argument that he failed to present. (Cf. Lankford v. Idaho, supra, 500 U.S. at pp. 122-124 [111 S.Ct. at pp. 1730-1732].) Thus the case here is distinguishable from Lanlford, and it more closely resembles Neder v. United States, supra, 527 U.S. 1, which applied harmless error analysis when the defendant was mistakenly required to present his evidence on the issue of materiality to the judge instead of the jury.
Concluding that the court’s failure to instruct on predatory acts is subject to harmless error analysis, we turn to the problem of discerning the applicable test of prejudice for constitutional error. The purpose of the SVPA is not punitive. (Kansas v. Hendricks (1997) 521 U.S. 346, 367, 369 [117 S.Ct. 2072, 2084-2085, 2085-2086, 138 L.Ed.2d 501]; Hubbart v. Superior Court (1999) 19 Cal.4th 1138, 1171-1177 [81 Cal.Rptr.2d 492, 969 P.2d 584].) Instead, its primary purpose is to protect the public from “a small but extremely dangerous group of sexually violent predators that have diagnosable mental disorders [that] can be identified while they are incarcerated.” (Stats. 1995, ch. 763, § 1.) Although treatment is a secondary objective (Hubbart v. Superior Court, supra, 19 Cal.4th at p. 1166), a defendant likely to commit future predatory acts can be committed even if his condition is not amenable to treatment. (See § 6606, subd. (b); Kansas v. Hendricks, supra, 521 U.S. at p. 366 [117 S.Ct. at p. 2084].)
Although the SVPA is a civil proceeding, its procedures have many of the trappings of a criminal proceeding. The probable cause hearing at issue here (§ 6602) is one example. In addition, the defendant is entitled to appointed counsel (§ 6603, subd. (a)), the trier of fact must find the defendant to be a sexually violent predator beyond a reasonable doubt (§ 6604), and a jury verdict must be unanimous (§ 6603, subd. (d)).
In sum, proceedings under the SVPA, in common with proceedings trader other civil commitment statutes, are civil proceedings with consequences comparable to a criminal conviction—involuntary commitment, often for an indefinite or renewable period, with associated damage to the defendant’s name and reputation. The United States Supreme Court and the California courts have pointed to these consequences in cases holding that the ordinary *1193burden of proof in civil cases, preponderance of the evidence, is constitutionally inadequate in cases involving involuntary commitment and, as we shall explain, the California courts have in turn relied on the burden of proof decisions to hold that federal constitutional error in civil commitment proceedings is reversible unless shown to be harmful beyond a reasonable doubt.
In Addington v. Texas (1979) 441 U.S. 418, 425 [99 S.Ct. 1804, 1809, 60 L.Ed.2d 323], the high court said: “This Court repeatedly has recognized that civil commitment for any purpose constitutes a significant deprivation of liberty that requires due process protection. [Citations.] Moreover, it is indisputable that involuntary commitment to a mental hospital after a finding of probable dangerousness to self or others can engender adverse social consequences to the individual. Whether we label this phenomena [sz'c] ‘stigma’ or choose to call it something else is less important than that we recognize that it can occur and that it can have a very significant impact on the individual.” The court concluded that the preponderance of the evidence standard of proof was constitutionally inadequate in a civil commitment proceeding. (Id. at p. 427 [99 S.Ct. at p. 1810].)
In People v. Burnick (1975) 14 Cal.3d 306 [121 Cal.Rptr. 488, 535 P.2d 352], we considered the burden of proof under the former Mentally Disordered Sex Offender Law, former section 6300 (MDSO Law), which provided for commitment of mentally disordered sex offenders. We said that “so drastic an impairment of the liberty and reputation of an individual must be justified by proof beyond a reasonable doubt.” (Burnick, at p. 310.) Four years later we decided Conservatorship of Roulet (1979) 23 Cal.3d 219 [152 Cal.Rptr. 425, 590 P.2d 1], which involved an action to extend a conservatorship (and thus to extend the confinement of the conservatee in a mental hospital) on the ground that the conservatee was gravely disabled. In requiring proof beyond a reasonable doubt, we said that “In People v. Burnick [citation], this court explicitly recognized that civil commitment to a mental hospital, despite its civil label, threatens a person’s liberty and dignity on as massive a scale as that traditionally associated with criminal prosecutions.” (Id. at p. 223.) The next year, in Conservatorship of Hofferber (1980) 28 Cal.3d 161 [167 Cal.Rptr. 854, 616 P.2d 836], we held that when a defendant is found incompetent to stand trial, his confinement as a person dangerous to others required proof beyond a reasonable doubt. “Involuntary confinement for mental illness or dangerousness,” Hofferber explained, “whether civil or criminal, involves loss of liberty and substantial stigma. Factfinding error must be minimized when such drastic consequences are at stake.” (Id. at p. 178.)
The burden of proof is not at issue here, but the reasoning of the cases discussing that issue applies equally to the question of the test for harmless *1194error, and is so applied in California decisions. In Conservatorship of Wilson (1982) 137 Cal.App.3d 132 [186 Cal.Rptr. 748], the Court of Appeal addressed the harmless error issue in a case involving erroneous instructions in a proceeding to commit a person as gravely disabled under the Lanterman-Petris-Short Act (§ 5000 et seq. (LPS Act)). Relying on Conservatorship of Roulet, supra, 23 Cal.3d 219, which involved the burden of proof under the LPS Act, Wilson reversed the commitment because the “error was not harmless beyond a reasonable doubt.” (Wilson, supra, 137 Cal.App.3d at p. 136.) Conservatorship of Early (1983) 35 Cal.3d 244, 255 [197 Cal.Rptr. 539, 673 P.2d 209], endorsed the holding in Wilson and held that in cases of civil commitment under the LPS Act constitutional error was reversible unless shown harmless beyond a reasonable doubt. People v. Lee (1980) 110 Cal.App.3d 774, 783 [168 Cal.Rptr. 231], applied the same standard in reviewing a commitment under the former MDSO Law.
An SVPA commitment unquestionably involves a deprivation of liberty, and a lasting stigma, equivalent to a commitment under the former MDSO law and probably more onerous than a commitment under the LPS Act. Although the commitment period under the SVPA is limited to two years, additional commitment proceedings can extend this period indefinitely. If the defendant has a mental disorder that is not susceptible to treatment, the proceedings may result in confinement for life. And whatever the consequences of being labeled as “gravely disabled,” or a “danger to himself or others” under the LPS Act, they are surely no greater than the consequences of a being labeled a “sexually violent predator.” Because the Chapman test (Chapman v. California, supra, 386 U.S. 18)—that federal constitutional error is reversible unless shown to be harmless beyond a reasonable doubt—is used for the review of federal constitutional error in civil commitment cases in California generally, that test necessarily governs review under the SVPA.
V
We agree with the Court of Appeal that the trial court’s error in failing to instruct the jury on the need to find a likelihood of fixture predatory acts was harmless beyond a reasonable doubt. Defendant was first adjudged a mentally disordered sex offender in 1979. He was paroled in 1984, violated parole, and was returned to prison. In 1985 defendant again violated parole and was convicted of annoying or molesting children. (Pen. Code, § 647.6.) He was paroled in August of 1987 and in October of that year was convicted of committing sodomy and lewd and lascivious acts on three minors, ages 12, 13, and 14. All three victims were strangers to defendant, which means that defendant’s acts were “predatory acts” as defined in section 6600, *1195subdivision (e). Defendant admitted to Dr. Goldberg, one of the examining psychiatrists, that he continued to have sexual fantasies about children. Thus there was ample evidence to show that defendant was likely to commit future violent sexual acts, and none to indicate that his victims would not include strangers, casual acquaintances, or persons cultivated for victimization. We therefore conclude that the trial court’s instructional error was harmless beyond a reasonable doubt.
Disposition
The judgment of the Court of Appeal is affirmed.
George, C. J., Werdegar, J., Chin, J., Brown, J., and Moreno, J., concurred.
Unless otherwise noted, all statutory citations are to the Welfare and Institutions Code.
Because there was a conflict in the Court of Appeal cases, the 2002 revision of CALJIC No. 4.19 gave the trial judge the option of instructing that the jury must find the defendant likely to engage in sexually violent predatory criminal behavior.
In 1998 the Legislature added two additional provisions referring to predatory acts. Section 6601.5, which provides for an urgency review when a defendant might be released before a section 6602 probable cause hearing can be held, requires the judge to determine whether the facts alleged in the petition “would constitute probable cause to believe that the [defendant] is likely to engage in sexually violent predatory criminal behavior upon his or her release.” Section 6602.5 prohibits state hospital placement unless there has been a determination that there is probable cause to believe the defendant will engage in sexually violent predatory behavior.
The definition of probable cause and other issues relating to the second section 6002 hearing are before this court in Cooley v. Superior Court (S094676).