Cooley v. Superior Court

Opinion MORENO, J. The Sexually Violent Predators Act (SVPA) (Welf. & Inst. Code, § 6600 et seq.) 1 allows for the involuntary commitment of certain convicted sex offenders, whose diagnosed mental disorders make them likely to reoffend if released at the end of their prison terms. The issues we address here arise out of the provision of the SVPA that requires the superior court to hold a “probable cause hearing” as the initial step in the judicial process required to civilly commit a potential sexually violent predator (SVP). (§ 6602, subd. (a).) Here, after a probable cause hearing was conducted pursuant to section 6602, subdivision (a), the superior court dismissed a petition filed by the District Attorney of Los Angeles County alleging that real party in interest, Paul Marentez, was an SVP. The Court of Appeal vacated the superior court’s order and set the matter for trial. We now resolve several issues related to the scope and substance of the probable cause determination required.by section 6602, subdivision (a), and we determine whether the Court of Appeal correctly decided that there was sufficient evidence to establish probable cause to believe that Marentez was likely to engage in sexually violent predatory criminal behavior upon his release. *236 We conclude that the section 6602 hearing requires the superior court to determine whether a reasonable person could entertain a strong suspicion that the petitioner has satisfied all the elements required for a civil commitment as an SVP, specifically, whether (1) the offender has been convicted of a qualifying sexually violent offense against at least two victims; (2) the offender has a diagnosable mental disorder; (3) the disorder makes it likely he or she will engage in sexually violent criminal conduct if released; and (4) this sexually violent criminal conduct will be predatory in nature. We also conclude that the phrase “likely to engage in sexually violent predatory criminal behavior upon . . . release,” as it appears in section 6602, subdivision (a), requires the superior court to determine whether the potential SVP presents a serious and well-founded risk of committing sexually violent criminal acts that will be of a predatory nature, and that the superior court must consider the offender’s amenability to treatment when making this determination. We further conclude that although we would ordinarily apply the standard of review used in appellate review of criminal preliminary hearings, in this case remand to the superior court is appropriate. The entire superior court proceeding was infected with error because the expert evidence presented by both the district attorney and Marentez failed to consider whether potential sexual violence would be “predatory” in nature and all the experts applied the wrong definition of “likely.” It would be improper, therefore, under these circumstances, to uphold any factual findings made by the superior court, or to independently review the record to make our own factual findings. We reverse the Court of Appeal’s judgment vacating the superior court’s order and remand with the instruction that the Court of Appeal should, in turn, remand the matter to the superior court in order to conduct a new probable cause hearing consistent with the views expressed herein. Facts In 1988, and again in 1994, Marentez was convicted of lewd conduct with a minor under the age of 14. (Pen. Code, § 288, subd. (a).) In the 1988 offense, he lured a three-year-old boy into a bathroom at a church and orally copulated him. At the time of his arrest, he also indicated that in 1980 he had digitally molested and orally copulated a six-year-old girl, although he was acquitted of these crimes by a jury. The 1994 offense involved his girlfriend’s six-year-old son. He took the boy swimming at the YMCA, where, in the shower, he touched the boy’s penis for several minutes and placed the child’s hand on his own penis for about a minute. He served prison terms for both crimes. *237 In 1998, before the scheduled date for his release on parole, the district attorney filed a petition for Marentez’s commitment under the SVPA. (§§ 6600, 6601, subd. (i).) In 2000, the superior court conducted a probable cause hearing. The district attorney presented evidence, including written evaluations and testimony by two experts, to the effect that Marentez was likely to commit sexually violent offenses in the future. Marentez presented rebuttal evidence, consisting of written evaluations and testimony by three experts, to the effect that he was not likely to reoffend. Because our resolution of this case turns on both the evidence presented by the experts at the hearing and on the contents of the superior court’s ruling we summarize both in some detail below. Testimony of the District Attorney's Experts The district attorney’s first expert, Barrie Glen, Ph.D., a licensed psychologist, prepared a written report concluding that Marentez had been convicted of a statutorily defined qualifying offense involving two victims, had the qualifying mental disorder pedophilia, and was likely to engage in sexually violent criminal behavior on his release. For the latter determination, she relied, in her initial report, on several risk factors. She noted his lack of insight into his actions, lack of empathy for others, and long history of substance abuse. She also observed that his last qualifying offense was committed while he was on parole, and that he lived in a residence where children were present, in violation of the terms of his parole. He denied committing sexual offenses, blaming his incarceration on drug use. He had never undergone sex offender treatment and had no positive plans for dealing with his sexual urges toward children or with his substance abuse. In her initial report, Glen did not define the term “likely” and did not give any specific probability of the likelihood of reoffense. In a supplemental report, Glen confirmed her prior assessment, adding that her conclusion was supported by Marentez’s score on a new actuarial instrument, the Static-99 test, which analyzes the potential for risk of reoffense on the basis of certain variables, including age, marital status, number of criminal offenses committed, gender of the victims, whether the victims were strangers or relatives, and whether the most recent sex offense involved the use of violence. Marentez’s score of 6 on the Static-99 test put him in the “high risk” category for reoffense, with a 52 percent minimum risk of reoffense within 15 years. Glen’s direct testimony at the probable cause hearing focused on disputing the judgment of Marentez’s expert witnesses. Specifically, she challenged *238 their reliance on various mitigating factors and their dependence on self-reporting by Marentez. She also reiterated the findings contained in her supplemental report, suggesting that “when you give [Marentez] the predictor instrument he comes out over 50 percent, and when you start looking at the other risk factors, he looks like he has enough of them [so that] he is well over 50 percent.” At the close of cross-examination she stated that “I didn’t rely on clinical judgment in this case” because “[h]e fits in very well with actuarial stuff.” When asked by the court to clarify her previous statement, she stated, “I am a clinician; so of course my clinical judgment is in there.” She explained that she had meant that it was not necessary to “go outside any sort of actuarial data and just completely rely on clinical judgment.” Jack Vognsen, Ph.D., a licensed psychologist, also prepared a written evaluation for the prosecution. He, too, emphasized Marentez’s record of sexual crimes against children and his child-related parole violations. Indeed, Marentez had explained that he took on the relationship with his latest victim “in order to prove to the court that the prohibition against being around children was unnecessary.” Vognsen observed that Marentez had “us[ed] his position as a trusting, cafegiving adult to make his victims cooperate in sexual conduct.” Although Marentez complained that he had not been offered sex offense counseling, he asserted that “I need no therapy.” Similarly, although he blamed substance abuse for his mishaps, he indicated no interest in drug abuse counseling, boasting “I am my only rehab.” He did not accept blame for his sexual offenses, either denying them or stating that he could not remember whether he had committed the offense because he was under the influence of alcohol and drugs. He showed little self-awareness: “Mr. Marentez does not appear to listen to himself, and he certainly does not expect his ideas and attitudes to be reasonable.” Vognsen reiterated his conclusions in a supplemental report, noting that they were further supported by Marentez’s high score on the Static-99 test. At the hearing, he testified that Marentez was unable to control his urges toward children, basing his opinion on past behavior and the fact that Marentez had violated parole conditions by being around children. He referred to the results of the Static-99 test, conceding that the actuarial test is “only moderately reliable as an overall instrument,” but observing that it is “the best we can come up with. . . . With a little bit of clinical judgment.” Using the Static-99 test and “[a] dash of clinical judgment,” he estimated Marentez’s likelihood of reoffense over a 15-year period at “52 to 55, 57 [percent], something like that.” *239 Marentez’s Expert Testimony Raymond E. Anderson, Ph.D.,. a licensed psychologist, appeared on behalf of Marentez. His written evaluation stated the “impression . . . that any sexually violent offense would be far less than 50% probable given all the data in this case.” In his view, Marentez was a “situational” rather than a “preferential” child sexual abuser; his offenses resulted from a “failure of control rather than from strong internal direction.” He observed, however, that Marentez showed a lack of empathy for his victims “similar to that seen in untreated sexual abusers.” Marentez should also “be more concerned with his involvement in [three] different [offenses] and the implication of this for his social judgment and functioning.” At the hearing, Anderson testified consistently with his written evaluation. He opined that Marentez had achieved greater self-restraint and perspective since his incarceration and was “less likely to reoffend than 50 percent with or without treatment.” He also testified extensively concerning the Static-99 test. He described it as the “best actuarial tool” for the purpose, but he also emphasized the test’s numerous shortcomings, i.e., its results are “very weak and inexact”; it has no known “base rate” for sexual recidivism; it “gives you a prediction of any sexual offense, not [just] a violent sexual offense”; it has overlapping “predictors” that might produce a falsely high score; and it has not been standardized against a population including Hispanics or Native Americans—Marentez’s ethnicity. He also stated that it was improper to supplement the variables contained in the Static-99 test with additional variables to adjust the actuarial prediction, because the additional variables may be intercorrelated with those already present in this test such that they would already be accounted for in the original prediction. He explained that “most people believe that the actuarial [test] should be used in tandem with clinical prediction.” With regard to Marentez, he observed, “I don’t think the two [actuarial prediction and clinical judgment] taken together can add up to his being more likely than not to reoffend.” William Vicary, M.D., a psychiatrist, also prepared a written evaluation for the defense. He found the case “relatively close,” noting that “reasonable experts could arrive at different conclusions.” He concluded that Marentez suffered from pedophilia and other mental disorders but was not more likely than not to reoffend as a result. Vicary utilized the RRASOR (rapid risk assessment of sex offender recidivisim) predictive test and found that Marentez scored a 3, representing a 37 percent chance of reoffehse within 10 years. He observed that Marentez was defensive with regard to his past criminal acts and “seemed to have an explanation or excuse for just about everything.” He was “less than forthright” about the molestation offenses *240 and also showed little remorse. He noted, however, that Marentez’s current incarceration appeared to have had “a significant impact” in that he was aware he was facing a possible life sentence for future offenses, that he had a stable employment history and prospects, and that his antisocial traits were likely to diminish with age. At the hearing, Vicary testified consistently with his evaluation that Marentez did not meet the criteria for SVP status. With regard to recidivism, he opined that Marentez “will always represent a significant risk of sexually reoffending. The question is, is there enough data to get us to more likely than not.” Based on all the data, including the results of the Static-99 test and other actuarial instruments, he disagreed with Glen and Vognsen that Marentez was more likely than not to reoffend. In his view, Marentez’s high score on the Static-99 test was a “significant element” and made it a “closer case,” but he regarded it as a screening device that must be adjusted by other factors. Marentez was now “older, wiser and scared,” his crimes were “obvious and impulsive and foolhardy,” making him less likely to offend than the “classic pedophile.” Rebecca L. Crandall, M.D., a psychiatrist, prepared a written evaluation concluding that Marentez did not suffer from a diagnosable mental illness and therefore did not meet the SVP criteria. She noted that he denied having sexual fantasies about children, was no longer using drugs, and displayed no antisocial behavior in prison. Crandall noted in her report, however, that Marentez also denied responsibility for two of the molestations, claiming to have no memory of the incident involving the three-year-old boy in the church bathroom, because he was intoxicated with alcohol and cocaine. At the hearing, Crandall reiterated her view that the evidence did not support a diagnosis of pedophilia. In evaluating Marentez, Crandall “basically concentrated on [her] area of expertise of psychiatric diagnosis.” Although she considered actuarial data, including the factors in the Static-99, she did not conduct any specific actuarial tests. Because of her failure to diagnose a mental disorder, she did not go on to assess whether Marentez was more likely than not to reoffend and offered no opinion on the point. Superior Court’s Probable Cause Ruling During argument, the court stated that it believed that the evidence established that Marentez suffered from pedophilia that affected his volitional or emotional capacity. It was concerned, however, with the validity of the expert opinions given their reliance on the Static-99 statistical instrument. The court observed that in previous cases, “State’s experts have *241 testified that the most relevant opinions are based upon clinical judgement informed by statistical evidence^]” The court noted that in this case “[o]ne of the State’s experts has testified that her opinion is based only on the statistical evidence, namely the Static-99, and the State’s second witness testified that his opinion was based on the results of Stat[ic]-99 basically with a ‘dash of clinical judgment.’ ” The court opined that “if a witness is relying on the statistical evidence to voice an opinion and the statistical evidence doesn’t with reasonable certainty predict an individual respondent’s recidivistic likelihood, then it seems to me that opinion is worthless for purposes of establishing probable cause.” The court stated that even if the Static-99 test were reliable enough, “it doesn’t take into account the dynamic factors, which are equally important in determining whether a person, even a pedophile or any other diagnosed mental disorder is . . . more likely than not ... to reoffend.” The court considered Anderson’s evidence as “impeaching the certainty of the state’s witnesses,” which were “based only on [actuarial test results].” The court also commented on the credibility of the experts. In regards to Glen’s testimony, the court noted that “she doesn’t routinely do forensic cases.” Of Vognsen, it observed: “His demeanor on the witness stand was . . . such as to lead me to believe that he really is not particularly confident of his answers to the questions about statistical evidence.” In applying a “ ‘dash of clinical judgment’ ” to the results of the Static-99 test, “[Vognsen] seems in his report and in his testimony to have reacted quite subjectively.” It also commented on Vognsen’s “apparent dislike” of Marentez. In contrast, the court noted that it “was extraordinarily impressed with [Anderson’s] knowledge of statistics and the method in which statistics—the Static-99 . . . and other tests—were devised.” The court then issued its probable cause ruling: “I am basing this decision on a reasonable cause standard . . . [and] on my view of the demeanor of the witnesses, the manner in which they testified, the certainty with which they had command of the underlying information upon which they based the decision, and also on the issues of the other factors which go into credibility. fl[] In this case I do believe that Dr. Glen in the end based her opinion upon Static-99 alone because that’s what she said. I do believe that while she did consider clinical matters, in the end, when asked she stated her opinion was based on the Static-99, which we know is not reliable enough according to accepted scientific principles to establish that this individual is more likely than not to commit a sexually violent offense. [^] With regard to Dr. Vognsen, I said earlier his testimony in court showed a lack of command of the statistical information as compared to Dr. Anderson’s. Again, he said his reliance was on Static-99. The import of his testimony was that Static-99 *242 was the basis of his opinion with a dash of clinical judgment. Again, Static-99, as I said, isn’t good enough. [|] . . . fl[] Dr. Anderson’s testimony with regard to the likelihood of reoffense was validated to some extent by the psychological testing that he gave. Overall, it’s the Court’s view that Dr. Anderson has far more experience based on evidence presented in administering this kind of testing and dealing with sexual offenders both on a clinical basis and on a research basis, that his testimony was carefully thought out and considered, which I did not find to be the case with the manner in which the two witnesses for [the district attorney] testified. [^|] Given that and the testimony of Dr. Vicary on the risk of reoffense, the Court does believe Dr. Vicary. This is a close case, but the Court does not find that the evidence supports a reasonable belief that Mr. Marentez is more likely than not to reoffend in a sexually violent manner[.]” On that basis the superior court dismissed the petition to commit Marentez as an SVP. The district attorney petitioned for a writ of mandate. A divided Court of Appeal determined that the superior court erred in “rejecting outright” evidence of the Static-99 results for purposes of Vognsen’s testimony. 2 According to the majority, the superior court improperly applied the standard of admitting new scientific evidence to a case involving psychiatric evaluation. (See People v. Kelly (1976) 17 Cal.3d 24 [130 Cal.Rptr. 144, 549 P.2d 1240]; Frye v. United States (D.C. Cir. 1923) 293 Fed. 1013.) 3 Based on its own independent review of the evidence it concluded that there was “ample rational basis for a reasonable person to assume the possibility that it is likely Marentez will engage in sexually violent predatory criminal behavior if released.” 4 In reaching this decision, the majority noted that it need not resolve the question of the correct definition of “likely” in this case because *243 “the evidence at bar supports a finding of probable cause under either a ‘substantial’ or ‘greater than 50 percent’ standard.” It issued the writ, ordering the superior court to reinstate the petition, enter a finding of probable cause, and set the matter for trial. The dissent disagreed, concluding that “[t]he assertion that the superior court judge excluded evidence of the Static-99 test ... is not supported by the record.” Instead, in determining that the district attorney’s experts were unpersuasive because they relied either solely or almost solely on the Static-99 test, the superior court “made specific factual findings that were dispositive on the issue of reoffense.” We granted review, and now turn to a determination of the issues in this case. Discussion Statutory Background The SVPA provides for the involuntary civil commitment of an offender immediately upon release from prison, for a two-year period, if the offender is found to be an SVP. The civil commitment can only commence if, after a trial, either a judge or a unanimous jury finds beyond a reasonable doubt that the person is an SVP (§§ 6600, 6601, 6603, 6604). 5 An SVP is defined in section 6600, subdivision (a)(1), as “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.” 6 In addition, we recently held that the predatory nature of criminal acts is an additional element required to be proven at trial in order to commit an SVP. (People v. Hurtado (2002) 28 Cal.4th 1179, 1186 [124 Cal.Rptr.2d 186, 52 P.3d 116] (Hurtado).) *244 The statute defines a “diagnosed mental disorder” as “a congenital or acquired condition affecting the emotional or volitional capacity that predisposes the person to the commission of criminal sexual acts in a degree constituting the person a menace to the health and safety of others.” (§ 6600, subd. (c).) The statute lists the convictions that qualify as “sexually violent” offenses when “committed by force, violence, duress, menace, or fear of immediate and unlawful bodily injury.” (Id., subd. (b).) 7 It also defines “predatory” acts as 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).) The trial, however, is the last stage of a complex administrative and judicial process to determine whether an offender should be civilly committed as an SVP. The offender must first be screened by the Department of Corrections at least six months before the scheduled release date from prison. (§ 6601, subd. (a).) If the department finds the offender is likely to be an SVP, he or she is referred to the Department of Mental Health for a “full evaluation.” (§ 6601, subd. (b).) The offender must then be evaluated by at least two mental health professionals designated by the Director of the Department of Mental Health. (§ 6601, subds. (c), (d).) Only if two evaluators concur that “the person has a diagnosed mental disorder so that he or she is likely to engage in acts of sexual violence without appropriate treatment and custody” (id., subd. (d)) can the Director of Mental Health forward a petition for civil commitment to the appropriate county—the county where the offender was convicted of the crime for which he or she is imprisoned. (Id., subds. (d)-(h); People v. Superior Court (Ghilotti) (2002) 27 Cal.4th 888, 909 [119 Cal.Rptr.2d 1, 44 P.3d 949] (Ghilotti).) Thereafter, designated counsel—in this case the district attorney—can file a petition for commitment in the superior court and initiate the judicial proceedings required by the act. (§ 6601, subd. (i).) Once the petition has been filed, it is reviewed by a superior court judge to determine whether it “states or contains sufficient facts that, if true, would constitute probable cause to believe that the individual named in the petition *245 is likely to engage in sexually violent predatory criminal behavior upon his or her release. If the judge determines that the petition, on its face, supports a finding of probable cause, the judge shall order that the person be detained in a secure facility until a hearing can be completed pursuant to Section 6602.” (§ 6601.5.) At the hearing, “[a] judge of the superior court shall review the petition and shall determine whether there is probable cause to believe that the individual named in the petition is likely to engage in sexually violent predatory criminal behavior upon his or her release. The person named in the petition shall be entitled to assistance of counsel at the probable cause hearing. ... If the judge determines there is not probable cause, he or she shall dismiss the petition and any person subject to parole shall report to parole. If the judge determines that there is probable cause, the judge shall order that the person remain in custody in a secure facility until a trial is completed and shall order that a trial be conducted to determine whether the person is, by reason of a diagnosed mental disorder, a danger to the health and safety of others in that the person is likely to engage in acts of sexual violence upon his or her release from the jurisdiction of the Department of Corrections or other secure facility.” (§ 6602, subd. (a).) Only if the superior court finds that there is probable cause, therefore, does the civil commitment process proceed beyond this initial judicial proceeding to trial. 8 Before reaching the merits, therefore, we must decide what this substantive determination of probable cause entails in the context of section 6602, subdivision (a). Specifically, we must (1) identify the elements which must be proved at the probable cause hearing; (2) define the standard of proof to be applied by the superior court; and (3) define the meaning of the phrase “likely to engage in sexually violent predatory criminal behavior” contained in section 6602, subdivision (a). Finally, we must also determine the standard of review to be applied to a superior court’s probable cause determination in the context of the SVPA. *246 Scope of Probable Cause Hearing We first resolve the issue, raised by the Court of Appeal and disputed by the parties, of the proper scope of this hearing. The outcome of the probable cause hearing determines whether or not an offender must proceed to a trial in which he or she might be civilly committed as a sexually violent predator. (§ 6602, subd. (a).) As previously noted, section 6602, subdivision (a), describes the determination to be made at the probable cause hearing as to whether there is “probable cause to believe that the individual ... is likely to engage in sexually violent predatory criminal behavior upon his or her release.” Section 6604, which describes the determination to be made at trial, requires that a court or jury find “beyond a reasonable doubt, the person is a sexually violent predator.” (Italics added.) A sexually violent predator is defined in section 6600, subdivision (a)(1), as “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.” In addition, as previously noted, we recently held that the predatory nature of criminal acts is an additional element required to be proven at trial in order to commit an SVP. (Hurtado, supra, 28 Cal.4th at p. 1186.) 9 The Legislature used different language to describe the ultimate determinations to be made at trial and at the probable cause hearing. Section 6604 describes the determination to be made at trial as whether the person is “a sexually violent predator,” whereas subdivision (a) of section 6602 describes the determination as whether the person is “likely to engage in sexually violent predatory criminal behavior upon his or her release.” The Court of Appeal apparently relied on this difference in the statutory language to conclude that “the only finding the court has to make” at the probable cause hearing “is whether there is probable cause to believe that the sexual predator is likely to engage in sexually violent criminal behavior upon release.” The Court of Appeal went on to state that “[t]he court is not required to make any findings on any of the other criteria of the Sexually Violent Predators Act at the probable cause hearing.” The district attorney adopts the Court of Appeal position here. *247 Marentez, on the other hand, argues that the scope of a probable cause finding encompasses all four of the elements contained in the definition of an SVP. He argues that this interpretation is implicit in the language contained in section 6602, subdivision (a), because only individuals with an identified sexual criminal history who have a mental disorder are statutorily defined as being “likely to engage in sexually violent predatory criminal behavior.” (§ 6602, subd. (a).) Contrary to the position taken by the Court of Appeal and adopted by the district attorney, we do not believe that the difference in language used in section 6602, subdivision (a) and section 6604 signifies an intention by the Legislature that the scope of the probable cause hearing should be more limited than the scope of the trial. Under the SVPA, an individual can only proceed to a trial if “the judge determines there is probable cause,” and the petition is dismissed if “there is not probable cause.” (§ 6602, subd. (a).) The probable cause hearing, therefore, is only a preliminary determination that cannot form the basis of a civil commitment; the ultimate determination of whether an individual can be committed as an SVP is made only at trial. (§ 6604.) For this reason, based on the structure of the SVPA, a section 6602 hearing is analogous to a preliminary hearing in a criminal case; both serve to “ ‘ “weed out groundless or unsupported charges . . . and to relieve the accused of the degradation and expense of a . . . trial.” ’ ” (Nienhouse v. Superior Court (1996) 42 Cal.App.4th 83, 91 [49 Cal.Rptr.2d 573].) Like a criminal preliminary hearing, the only purpose of the probable cause hearing is to test the sufficiency of the evidence supporting the SVPA petition. (Hurtado, supra, 28 Cal.4th at p. 1186.) Given this purpose, we can discern no reason why the superior court at the probable cause hearing would test the sufficiency of only one of the elements required for civil commitment as an SVP at trial. We do not believe that the Legislature intended such a result. Rather, we interpret section 6602, subdivision (a) in light of the language used and the purpose of the probable cause hearing within the structure of the SVPA, to require that the probable cause hearing encompass all four of the elements also required for the ultimate determination at trial. We proceed by examining the language contained in section 6602, subdivision (a). (Lungren v. Deukmejian (1988) 45 Cal.3d 727, 735 [248 Cal.Rptr. 115, 755 P.2d 299].) Pursuant to this section, the superior court must determine at the probable cause hearing whether the offender is “likely to engage in sexually violent predatory criminal behavior upon his or her release.” (§ 6602, subd. (a), italics added.) Viewing this language in isolation, it is not clear how the Legislature intended that the superior court make *248 this determination. For purposes of the determination to be made at trial, however, an SVP is an individual with a qualifying sexual offense and 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), italics added.) We can infer from the use of this similar language that the Legislature intended the determination of dangerousness at the probable cause hearing to be the same as at the subsequent trial. In fact, the district attorney does not suggest otherwise. The district attorney suggests, instead, that the probable cause hearing is limited solely to a determination of the “likely” element. The district attorney’s position relies on the fact that the existence of a mental condition and dangerousness, manifested in the likelihood of sexual criminal behavior, are different elements contained in the definition of an SVP that must be proved at trial. (§§ 6600, subd. (a), 6604.) The district attorney suggests, therefore, that these elements can be separated, and that the probable cause hearing is limited solely to a determination of dangerousness. This argument ignores the fact that in the definition of an SVP contained in section 6600, subdivision (a), a finding of “likely [to] engage in sexually violent criminal behavior” is expressly dependent on the existence of a statutorily defined mental disorder: “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.” (Ibid., italics added.) Based on the statutory definition of an SVP, therefore, at trial, the finder of fact can only find that an offender is likely to engage in sexually violent criminal behavior if the offender suffers from a mental disorder. Applying the rule of statutory construction that “ ‘provisions relating to the same subject matter must be harmonized to the extent possible’ ” (Lakin v. Watkins Associated Industries (1993) 6 Cal.4th 644, 659 [25 Cal.Rptr.2d 109, 863 P.2d 179]), we find that a finding of the existence of a mental disorder is, therefore, also implicit in the determination of dangerousness to be made at the probable cause hearing. We can conclude, therefore, that the probable cause hearing necessarily encompasses at least two of the elements required at trial: a finding by the superior court of a statutorily defined diagnosed mental disorder and a finding of future dangerousness—whether a person is likely to commit sexually violent predatory behavior upon release. 10 We rely on the plain language of section 6602, subdivision (a) to conclude that the superior court must also determine whether there is *249 probable cause to believe it likely that future acts of sexual violence will be “predatory.” This subdivision specifically refers to “sexually violent predatory criminal behavior.” (Ibid., italics added.) Furthermore, as previously noted, the term “predatory” is expressly defined in section 6600, subdivision (e). To ignore the term, therefore, would violate the maxim of statutory construction that “ ‘[c]ourts should give meaning to every word of a statute if possible, and should avoid a construction making any word surplusage.’ [Citation.]” (Reno v. Baird (1998) 18 Cal.4th 640, 658 [76 Cal.Rptr.2d 499, 957 P.2d 1333].) Our conclusion reiterates our previous interpretation of the plain meaning of the language of the SVPA probable cause hearing provision. (See People v. Torres (2001) 25 Cal.4th 680, 686 [106 Cal.Rptr.2d 824, 22 P.3d 871] [“language [of section 6602, subdivision (a)] 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’ ” (italics added)]; Hurtado, supra, 28 Cal.4th at p. 1183.) We further conclude that the fourth element of an SVP determination— whether the offender has been convicted of at least one qualifying offense (§ 6600, subd. (a)) falls within the scope of the probable cause hearing. In reaching our conclusion, we recognize that the language of section 6602, subdivision (a) contains no express reference to the offender’s qualifying conviction. Given that the purpose of the probable cause hearing is to test the sufficiency of evidence for civil commitment, however, we can discern no reason why the superior court at the probable cause hearing would test the sufficiency of all the elements except the existence of a qualifying conviction when an individual cannot be found to be an SVP at trial without this element. We note that the statute expressly provides that the existence of a qualifying offense can be proved by documentary evidence, so that proof of this element is unlikely to be a subject of dispute. (§ 6600, subd. (a)(3).) The probable cause determination, however, is the first judicial hearing at which the offender can point out any error related to this documentary evidence. In light of the fact that this element can be easily verified, excluding this determination from the scope of the probable cause hearing would lead to the absurd result that an individual could potentially be kept in custody pending an eventual trial, even though he or she did not meet this basic requirement for civil commitment. (§ 6602, subd. (a).) Therefore, based on the purpose of section 6602, subdivision (a), within the structure of the SVPA, we conclude that the probable cause hearing must also encompass a judicial verification of the existence of at least one qualifying offense as defined in section 6600, subdivision (a). *250 In sum, we hold that the probable cause determination, like the ultimate determination to be made at trial, encompasses all four of the elements contained in the definition of an SVP. Burden of Proof Next, we must determine what the appropriate burden of proof is for a probable cause hearing under the SVPA. In reaching their decisions in this case, both the superior court and the Court of Appeal applied the “probable cause” standard as used in the criminal preliminary hearing as the burden of proof. Both parties appeared to agree in the proceedings below that the “probable cause” standard was the proper burden of proof. It was not until the petition for rehearing in the Court of Appeal that Marentez claimed, as he does now, that because proceedings under the SVPA are “civil” in nature it is “inappropriate” to apply a “criminal” standard at this hearing. He now argues that the burden of proof at the probable cause hearing is governed by section 115 of the Evidence Code, and that, therefore, the higher burden of “preponderance of evidence” applies. He also argues that we should “read” a “preponderance of evidence” burden of proof into section 6602, subdivision (a) in order to avoid an equal protection problem. 11 Marentez is correct that because SVPA proceedings are civil in nature (Hubbart v. Superior Court (1999) 19 Cal.4th 1138, 1172 [81 Cal.Rptr.2d 492, 969 P.2d 584] (Hubbart); People v. Superior Court (Preciado) (2001) 87 Cal.App.4th 1122, 1128 [105 Cal.Rptr.2d 159]), section 115 of the Evidence Code applies here. The relevant part of this section states: “Except as otherwise provided by law, the burden of proof requires proof by preponderance of evidence.” (Evid. Code, § 115, italics added.) 12 Therefore, although this section of the Evidence Code requires that the “preponderance of *251 evidence” burden of proof be the default burden in a civil case, the section also allows the Legislature to specify a different burden of proof in certain civil proceedings. Applying Evidence Code section 115 here, we find that Welfare and Institutions Code section 6602, subdivision (a) clearly comes within the “otherwise provided” caveat. This provision of the SVPA states that a superior court judge “shall determine whether there is probable cause to believe that the individual named in the petition is likely to engage in sexually violent predatory criminal behavior upon his or her release.” (Welf. & Inst. Code, § 6602, subd. (a).) The text of this section refers to the hearing as a “probable cause hearing,” and states that the civil commitment can only proceed to the trial stage if there is “probable cause.” (Ibid.) The Law Revision Commission comment to Evidence Code section 115 even states that the “ ‘sufficient cause’ ” burden—analogous to the probable cause burden—is an example of a burden of proof prescribed by law that might be required instead of the preponderance of evidence burden. (Cal. Law Revision Com. com., Deering’s Ann. Evid. Code (1986 ed.) foil. § 115, p. 14.) Although the term “probable cause” is not defined in the SVPA, “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].) This court has stated in the felony preliminary hearing context that “ ‘ “[p]robable cause is shown if a man of ordinary caution or prudence would be led to believe and conscientiously entertain a strong suspicion of the guilt of the accused.” ’ [Citations.]” (Rideout v. Superior Court (1967) 67 Cal.2d 471, 474 [62 Cal.Rptr. 581, 432 P.2d 197]; People v. Uhlemann (1973) 9 Cal.3d 662, 667 [108 Cal.Rptr. 657, 511 P.2d 609]; People v. Slaughter (1984) 35 Cal.3d 629, 636 [200 Cal.Rptr. 448, 677 P.2d 854] (Slaughter).) In making the determination of probable cause, the magistrates do not themselves decide whether the defendant is guilty. (See Slaughter, supra, 35 Cal.3d at p. 637; Uhlemann, supra, 9 Cal.3d at p. 667.) Rather, they simply decide whether a reasonable person could harbor a strong suspicion of the defendant’s guilt. In doing so, they may “weigh the evidence, resolve conflicts, and give or withhold credence to particular witnesses.” (Uhlemann, supra, 9 Cal.3d at p. 668.) But the proceeding is not a trial: if the magistrate forms a personal opinion regarding the defendant’s guilt, it is of no legal significance. (Slaughter, supra, 35 Cal.3d at p. 637.) In sum, the magistrate’s role is limited to determining whether a reasonable *252 person could harbor a strong suspicion of the defendant’s guilt, i.e., whether such a person could reasonably weigh the evidence, resolve conflicts, and give or withhold credence to particular witnesses in favor of harboring such a suspicion. We assume, therefore, that the Legislature, by using the term “probable cause” in section 6602, subdivision (a), intended an analogous definition and application of this term in the context of this civil commitment scheme. (See Hurtado, supra, 28 Cal.4th at p. 1189.) We conclude, therefore, that a determination of probable cause by a superior court judge under the SVPA entails a decision whether a reasonable person could entertain a strong suspicion that the offender is an SVP. In reaching this conclusion, we reject Marentez’s suggestion that it is “inappropriate” for the Legislature to utilize a burden of proof in a civil commitment scheme because it has also been developed in the criminal context. It is correct that the SVPA is a civil commitment scheme, and therefore, we should not construe the consequence of commitment as “criminal” or “punitive” in nature. (Hubbart, supra, 19 Cal.4th at p. 1172; Kansas v. Hendricks (1997) 521 U.S. 346, 369 [117 S.Ct. 2072, 138 L.Ed.2d 501].) There is a difference, however, between characterizing a civil commitment as a criminal or punitive sanction and borrowing certain procedural protections or standards previously established and developed under criminal law. As we have noted: “the use of procedural safeguards traditionally found in criminal trials [does] not mean that commitment proceedings [are] penal in nature.” (Hubbart, supra, 19 Cal.4th at p. 1174, in. 33; Hendricks, supra, 521 U.S. at pp. 364-365 [117 S.Ct. at pp. 2083-2084].) We also reject Marentez’s equal protection challenge to the employment of the “probable cause” burden of proof in section 6602 hearings—which we assume to be an invocation of the “familiar rule of construction that statutes should be interpreted in a manner which avoids constitutional difficulties.” (Hutnick v. United States Fidelity & Guaranty Co. (1988) 47 Cal.3d 456, 466 [253 Cal.Rptr. 236, 763 P.2d 1326].) Marentez notes that the preponderance of evidence standard is applied under certain provisions of the Lanterman-Petris-Short Act (LPS Act). (§ 5150; In re Azzarella (1989) 207 Cal.App.3d 1240 [254 Cal.Rptr. 922] [preponderance of evidence is the burden of proof at habeas corpus proceeding challenging 14-day intensive commitment]; In re Lois M. (1989) 214 Cal.App.3d 1036 [263 Cal.Rptr. 100] [preponderance of evidence is the burden of proof at habeas corpus proceeding challenging 30-day temporary conservatorship].) He argues that a habeas corpus proceeding under the LPS Act is the “functional equivalent to a probable cause hearing” under the *253 SVPA because both are pretrial proceedings that determine whether or not an individual can be confined based on the fact that the person is a danger to himself or others because of a mental disorder. He argues, therefore, that the equal protection clauses of the United States and California Constitutions compel the same procedural safeguards in both types of proceedings. “ ‘The concept of the equal protection of the laws compels recognition of the proposition that persons similarly situated with respect to the legitimate purpose of the law receive like treatment.’ ” (In re Gary W. (1971) 5 Cal.3d 296, 303 [96 Cal.Rptr. 1, 486 P.2d 1201].) “The first prerequisite to a meritorious claim under the equal protection clause is a showing that the state has adopted a classification that affects two or more similarly situated groups in an unequal manner.” (In re Eric J. (1979) 25 Cal.3d 522, 530 [159 Cal.Rptr. 317, 601 P.2d 549]; People v. Buffington (1999) 74 Cal.App.4th 1149, 1154 [88 Cal.Rptr.2d 696]; People v. Gibson (1988) 204 Cal.App.3d 1425, 1436 [252 Cal.Rptr. 56].) This initial inquiry is not whether persons are similarly situated for all purposes, but “whether they are similarly situated for purposes of the law challenged.” (Gibson, supra, 204 Cal.App.3d at p. 1438.) Our equal protection inquiry begins, therefore, by determining whether persons subject to SVPA probable cause hearing are similarly situated to persons seeking habeas corpus relief from 14- or 30-day civil commitments under the cited provisions of the LPS Act. 13 Although the LPS Act and the SVPA are both civil commitment statutes, the Legislature enacted these schemes with different purposes in mind. The LPS Act is a comprehensive scheme designed to address a variety of circumstances in which a member of the general population may need to be evaluated or treated for different lengths of time. (§5150 [short-term emergency evaluation]; § 5250 [intensive 14-day treatment]; § 5300 [180-day commitment for the imminently dangerous]; § 5260 [extended commitment for the suicidal]; § 5350 [30-day temporary conservatorship or one year conservatorship for the gravely disabled].) In contrast, the SVPA narrowly targets “a small but extremely dangerous group of sexually violent predators that have diagnosable mental disorders [who] can be identified while they are incarcerated.” (Stats. 1995, ch. 763, § 1, p. 5921.) A stated purpose of the LPS Act is to provide “prompt evaluation and treatment of persons [from the general population] with serious mental disorders.” (§ 5001, subd. (b).) Such a purpose reflects the unfortunate reality that mental illness in its most acute form can pose a danger to the *254 individuals themselves or others that requires immediate attention. To achieve this purpose, a number of LPS Act provisions allow a person to be removed from the general population in order to be civilly committed based on a probable cause determination made by a mental health or law enforcement professional, and then to challenge the civil commitment within a reasonable time afterwards. Under the provisions of the LPS Act cited by Marentez (§§ 5350, 5352), an individual seeking habeas corpus relief has already been civilly committed for a specified period—14 or 30 days—and is challenging the commitment after the fact. (§§ 5275, 5276.) In contrast, under the SVPA, there is no comparable need to commit individuals for evaluation and treatment on an expedited basis. The SVPA scheme, therefore, allows for greater procedural safeguards, including both a probable cause hearing and trial, before the actual civil commitment. Although an individual may “remain in [pretrial] custody in a secure facility” after a finding of probable cause (§ 6602, subd. (a)), the actual civil commitment will only commence after an SVPA determination at the subsequent mandated trial. (§ 6604.) For this reason, the LPS Act habeas corpus proceedings and the SVPA probable cause hearing are not “functional equivalents” and persons utilizing these procedural protections are not “similarly situated.” As Marentez’s claim does not satisfy this preliminary requirement, his equal protection argument must fail. (See People v. Hubbart (2001) 88 Cal.App.4th 1202, 1221 [106 Cal.Rptr.2d 490].) There is no equal protection concern, therefore, that would bar application of the “probable cause” burden of proof at the hearing required by section 6602, subdivision (a). We hold, therefore, that the Legislature intended that the superior court apply the “probable cause” burden of proof at the hearing required by section 6602, subdivision (a), and that such burden does not violate the equal protection clauses of the California or United States Constitution. Definition of “Likely ” We must still, however, define the word “likely” in the context of the phrase “likely to engage in sexually violent predatory criminal behavior upon his or her release” as stated in section 6602, subdivision (a) (italics added). We made clear in Ghilotti that the determination of likelihood of future dangerousness was an element that must be proved in addition to the existence of mental disorder in order to commit an individual as an SVP. (Ghilotti, supra, 27 Cal.4th at pp. 920-921.) In that case, we defined the phrase “likely to engage in acts of sexual violence without appropriate *255 treatment and custody” (italics added) in the context of the psychological evaluations required to initiate commitment proceedings under the SVPA under section 6601, subdivision (d). We held that “an evaluator applying this standard must conclude that the person is ‘likely’ to reoffend if, because of a current mental disorder which makes it difficult or impossible to restrain sexually violent behavior, the person presents a substantial danger, that is, a serious and well-founded risk, that he or she will commit such crimes if free in the community.” (Ghilotti, supra, at p. 922.) We specifically rejected the argument that, in order to meet the “likely” standard, an evaluator must determine there is a “better than even [i.e., more likely than not] chance of new criminal sexual violence” (id. at p. 895). We declined to decide, however, whether the word “likely,” as used in section 6601, subdivision (d), is “similar to the statutory standard for final commitment [citations].” (Ghilotti, supra, 27 Cal.4th at p. 925, fn. 15, italics omitted.) 14 By implication, therefore, we declined to decide whether “likely” had a similar meaning in the context of the SVPA probable cause hearing. The district attorney argues that “likely” has the same meaning at the probable cause hearing as it does in the provision of the SVPA outlining the requirements of the psychological evaluations upon which the initial petition for commitment must be based. Marentez offers no substantive argument in response. 15 We agree with the district attorney. We find no support in the statutory scheme or the legislative history for the notion that the Legislature intended a different definition of “likely” to apply at the probable cause determination. “[A] word or phrase will be given the same meaning each time it appears in a statute . . . .” (Steketee v. Lintz, Williams & Rothberg (1985) 38 Cal.3d 46, 53 [210 Cal.Rptr. 781, 694 P.2d 1153]; Chandis Securities Co. v. City of Dana Point (1996) 52 Cal.App.4th 475, 486 [60 Cal.Rptr.2d 481].) Furthermore, the determination at the probable cause hearing is based on the petition filed by designated counsel, which is, in turn, necessarily based on the two concurring psychological evaluations required by section 6601. (§ 6602, subd. (a) [“A judge of the superior court shall review the petition and shall determine whether there is probable cause . . . .” (Italics added.)].) It would be illogical and impractical, therefore, for the psychological experts who based their conclusion on *256 the Ghilotti definition of the term in their initial evaluations supporting the petition for commitment, to offer conclusions based on a different meaning at the probable cause hearing. We conclude, therefore, that at the probable cause hearing the superior court must find probable cause to believe that a potential SVP presents a serious and well-founded risk of committing sexually violent criminal acts that will be of a predatory nature. We also conclude that a determination of the likelihood of future dangerousness at the probable cause hearing if such evidence has been presented must also take into account the potential SVP’s amenability to voluntary treatment upon release. In Ghilotti, we noted that the likelihood determination to be made by the psychological evaluators, required by section 6601, subdivision (d), was qualified by the phrase “without appropriate treatment and custody.''’ (Ghilotti, supra, 27 Cal.4th at p. 925.) We interpreted this phrase to mean that “the need for treatment and the need for custody are not always one and the same,” and therefore, the evaluators must determine whether “the person presents a substantial danger of reoffense if released without conditions, or whether instead he is safe only if restrained, supervised, and treated involuntarily [in] custody.” (Id. at pp. 926-927.) We noted that this qualifying factor to be applied to the dangerousness determination was required by both “section 6601, subdivision (d)” and “the SVPA in general.” (Id. at p. 926.) Section 6602, subdivision (a) does not expressly include the qualifying phrase “without appropriate treatment and custody” in defining the likelihood of “sexually violent predatory criminal behavior upon . . . release.” (Italics added.) Both phrases, however, similarly imply that the issue is whether the person is likely to reoffend without the confinement and involuntary treatment provided under the SVPA. Furthermore, as we observed with regard to the meaning of “likely,” it would be illogical, given the structure of the SVPA, to require psychological evaluators to provide a different estimation of the likelihood of reoffense at the probable cause hearing than that which formed the basis of the commitment petition. We hold, therefore, that the superior court at the probable cause hearing must also consider any evidence of the offender’s amenability to voluntary treatment in determining whether the potential SVP poses a serious and well-founded risk of committing sexually violent predatory criminal acts upon release. Standard of Review Having delineated the various substantive determinations to be made at a hearing required by section 6602, subdivision (a), we must decide, before reaching the merits of this case, the appropriate standard for reviewing the superior court’s ruling. *257 We agree with the Court of Appeal and the district attorney that the standard of review for preliminary hearings in criminal cases should also apply to review of probable cause determinations in SVP cases. We note that neither the statute itself nor the legislative history provides any guidance on this issue. In delineating the requirements of the SVPA probable cause hearing, however, we have concluded that the burden of proof is analogous to the burden imposed on the prosecution at a criminal preliminary hearing, and that, as in a criminal preliminary hearing, all the elements to be proved at trial must also be proved at the SVPA’s probable cause hearing. Our conclusion, in essence, models the requirements of the SVPA probable cause hearing on those contained in a criminal preliminary hearing. This similarity leads to the conclusion that we should also apply the same standard of review found in such hearings. (Cf. People v. Mercer (1999) 70 Cal.App.4th 463, 465-466 [82 Cal.Rptr.2d 723] [applying the same standard of review used in criminal cases following conviction to review of commitment order under the SVPA].) The resolution of mixed questions of law and fact, like probable cause, usually is examined independently (People v. Ashmus (1991) 54 Cal.3d 932, 1006-1007 [2 Cal.Rptr.2d 112, 820 P.2d 214]), and the resolution of a question of fact, like any such question underlying probable cause, always is examined for substantial evidence (e.g., People v. Waidla (2000) 22 Cal.4th 690, 730 [94 Cal.Rptr.2d 396, 996 P.2d 46]). Therefore, when reviewing a probable cause determination made pursuant to section 6602, subdivision (a), “[t]he character of judicial review . . . depends on whether the [superior court] has exercised [its] power to render findings of fact. If [it] has made findings, those findings are conclusive if supported by substantial evidence. [Citations.] If [it] has not rendered findings, however, the reviewing court cannot assume that [it] has resolved factual disputes or passed upon the credibility of witnesses. A dismissal unsupported by findings therefore receives the independent scrutiny appropriate for review of questions of law.” (Slaughter, supra, 35 Cal.3d at p. 638.) As we have previously noted in describing the burden of proof required at a probable cause hearing, “ ‘the [superior court] may weigh the evidence, resolve conflicts, and give or withhold credence to particular witnesses.’ ” (Slaughter, supra, 35 Cal.3d at p. 637.) In performing its role at the probable cause hearing, therefore, the superior court may evaluate the validity of any evidence presented by an expert, as well as judge the credibility of any expert witness who testifies at the hearing. Any credibility determination to be made at the probable cause stage, however, whether in a civil or criminal proceeding, is a gross and unrefined one. The superior court should not find an absence of probable cause simply because it finds the *258 defense witnesses slightly more persuasive than the prosecution witnesses. Rather, to reject the prosecution evidence at the probable cause stage, either the evidence presented must be inherently implausible, the witnesses must be conclusively impeached, or the demeanor of the witnesses must be so poor that no reasonable person would find them credible. Thus, if the prosecution presents evidence a reasonable person could accept over that presented by the defense, probable cause should be found. The superior court may not substitute its own personal belief as to the ultimate determination to be made at trial for that of a reasonable person evaluating the evidence. (Ibid.) These adjudicatory determinations are considered findings of fact for purposes of a probable cause determination: “ ‘The credibility of witnesses ... is a question of fact within the province of the committing magistrate to determine, and ... an appellate court may [not] substitute its judgment as to such question for that of a magistrate. [Citations.]’ ” (Jones v. Superior Court (1971) 4 Cal.3d 660, 667 [94 Cal.Rptr. 289, 483 P.2d 1241], quoting De Mond v. Superior Court (1962) 57 Cal.2d 340, 345 [19 Cal.Rptr. 313, 368 P.2d 865].) The nature of an appellate court’s review of the lower court’s probable cause determination also does not vary according to the type of evidence—whether it be lay or expert testimony. A reviewing court must still draw every inference in favor of the superior court’s factual findings and cannot substitute its judgment as to the credibility of the witnesses or the weight of the evidence over that of the magistrate. (See People v. Laiwa (1983) 34 Cal.3d 711, 718 [195 Cal.Rptr. 503, 669 P.2d 1278].) Analysis of Probable Cause Determination As the character of judicial review of this probable cause determination ordinarily depends on whether the superior court “exercised [its] power to render findings of fact” (Slaughter, supra, 35 Cal.3d at p. 638), we begin with a review of the superior court’s probable cause ruling to determine whether it did, in fact, make any conclusive factual findings to support its probable cause determination. In this case, the superior court properly exercised its adjudicatory functions to render the findings of fact that the opinions of the district attorney’s experts lacked persuasiveness. The superior court found Glen to lack persuasiveness because she relied solely on the Static-99 test. The superior court also found Vognsen’s testimony to be unpersuasive because he demonstrated a lack of command of the statistical information and relied almost completely on the Static-99 test. The Court of Appeal concluded that the superior court erred in this case by rejecting the testimony of Vognsen. The Court of Appeal majority *259 determined that the superior court improperly excluded the results of the Static-99 test for purposes of Vognsen’s testimony under the Kelly standard for admissibility of new scientific evidence. (See People v. Kelly, supra, 17 Cal.3d 24.) According to the Court of Appeal, the superior court erroneously ruled that a risk of reoffense derived from an actuarial instrument could not be used as the basis of an expert opinion. Our review of the record persuades us that the superior court did not, in fact, expressly purport to exclude such evidence; nor did it do so impliedly. Indeed, although it was critical of the reliability of the Static-99 test as a predictive tool—at least when used on its own, without adjustment by the use of clinical factors—it entertained testimony by all of the experts concerning the Static-99 test. Thus we conclude that the superior court made a proper factual finding as to the persuasiveness of both Glen and Vognsen based, in part, on their overreliance on this statistical instrument. The court also made the factual finding that the testimony given by Marentez’s witnesses was more credible and persuasive than that given by the district attorney’s experts. The court expressly noted that Anderson’s judgment was partially validated by the psychological testing, and that given his vast clinical and research experience, his testimony was well thought out and considered when compared to the district attorney’s two experts. The superior court also found the testimony of Marentez’s other witness, Vicary, to be persuasive. Based on these factual findings, the court concluded that there was no reasonable or probable cause, and dismissed the petition. Ordinarily, we would simply review the superior court’s factual findings for substantial evidence in order to determine whether to uphold the court’s probable cause determination. (Slaughter, supra, 35 Cal.3d at p. 638.) Our review is made more complicated, however, by the fact that although the superior court’s findings of credibility are supported by the evidence, the superior court’s determination was based on the improper definition of the word “likely” in the context of the phrase “likely to engage in sexually violent predatory criminal behavior upon . . . release” as stated in section 6602, subdivision (a) (italics added). In announcing its ruling, the court stated that it did not believe that “the evidence supports a reasonable belief that Mr. Marentez is more likely than not to reoffend in a sexually violent manner.” (Italics added.) Furthermore, the court did not reach the issue of whether there was probable cause to believe that violent criminal behavior in the future would be “predatory” in nature, as that term is defined in section 6600, subdivision (e). 16 Moreover, not only did the superior court’s actual ruling apply the wrong “likely standard,” but all the expert evidence introduced by both parties *260 rested on the assumption that the determination to be made was whether Marentez was more likely than not to reoffend. None of the experts were ever asked to consider the definition of “likely” that we conclude applies in the context of a probable cause hearing—whether there is a substantial danger, that is, a serious and well-founded risk, that Marentez is likely to reoffend. Nor did they consider whether their statistical predictions or clinical judgment would be altered by Marentez’s amenability to voluntary treatment were he to be released. Finally, none of the experts testified as to whether they believed Marentez was likely to commit sexually predatory criminal acts in the future—a requirement that might also affect their statistical predictions and clinical judgment. In essence, therefore, the probable cause hearing can be characterized as a battle of experts, some more credible than others, but all failing to consider all of the required elements, and all applying the wrong definition of the only element in dispute. We conclude, therefore, that the entire proceeding was infected with error. While we could undertake an independent review of the record and make our own determinations as to the existence of the SVP elements, such an inquiry would necessarily require some speculation in an area of highly specialized expert testimony. We note that superior court characterized its ruling as a “close case” based on the more-likely-than-not standard—a standard higher than that which we hold is applicable at the probable cause hearing. Indeed, Marentez’s experts highlighted several factors that indicate there is some risk that Marentez will reoffend. In fact, one of Marentez’s experts, Vicary, stated at the probable cause hearing that Marentez “will always represent a significant risk of sexually reoffending.” When making this statement, however, Vicary was not applying the standard that we articulate today. Given the considerable reliance by most of the experts on complex statistical predictors, we do not believe it wise to extrapolate from this evidence that that there is probable cause to believe that Marentez would pose a substantial danger or serious well-founded risk of sexually predatory criminal behavior upon release. It is also unclear from the record whether there is probable cause to believe Marentez is likely to commit “predatory” offenses in the future. Anderson, Marentez’s expert, did suggest that Marentez was a “situational offender,” but none of the experts specifically addressed the “predatory” element. The record suggests that Marentez’s offenses may have been related to his substance abuse, and thus it is unclear whether his history of *261 past offenses necessarily establishes a clear pattern of “predatory” over “situational” or improper “familial” sexual conduct. (Cf. Hurtado, supra, 28 Cal.4th at p. 1194 [“All three victims were strangers to defendant, which means that defendant’s acts were ‘predatory acts’ as defined in section 6600, subdivision (e).”].) Because of the incorrect application of the correct definition of “likely” by both parties and the trial court, and because the parties did not address whether Marentez’s future sexual violence would be “predatory” in nature, we believe that an independent review of the record by this court would be improper, and that remand to the superior court is appropriate. (See Ghilotti, supra, 27 Cal.4th at p. 895.) Conclusion Under these circumstances, we conclude that we must reverse the judgment of the Court of Appeal vacating the superior court’s order. We remand the cause to the Court of Appeal with the instruction that it should, in turn, remand the matter to the superior court in order to conduct a new probable cause hearing consistent with the views expressed herein. George, C. J., Kennard, J., Baxter, J., Werdegar, J., Chin, J., and Brown, J., concurred. Petitioner’s petition for a rehearing was denied January 15, 2003, and the opinion was modified to read as printed above. Footnotes [1] All further unlabeled statutory references are to the Welfare and Institutions Code. [2] The Court of Appeal majority concluded, however, that the superior court was within its discretion in excluding Glen’s testimony based on the court’s factual finding that no clinical judgment was used to supplement the predictive results of the Static-99 test. [3] Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993) 509 U.S. 579 [113 S.Ct. 2786, 125 L.Ed.2d 469] concluded that Frye v. United States, supra, 293 F. 1013, was overruled by the adoption of the Federal Rules of Evidence and replaced with a lower standard for scientific testimony. (Daubert, supra, 509 U.S. at pp. 587-590 [113 S.Ct. at pp. 2793-2795].) Daubert, however, does not displace People v. Kelly, supra, 17 Cal.3d 24, which construed the California Evidence Code in adopting its standard for admissibility. Therefore, although referred to as the Kelly/Frye test by the Court of Appeal in this case, we have previously noted that the foundational requirement for admission of new scientific evidence in California should now be referred to as the Kelly test or rule. (People v. Soto (1999) 21 Cal.4th 512, 515, fn. 3 [88 Cal.Rptr.2d 34, 981 P.2d 958].) [4] The court also noted, although it was not an issue raised by either party, that “at the probable cause hearing the only finding that the court has to make is whether there is probable cause to believe that the alleged sexually violent predator is likely to engage in sexually violent criminal behavior upon release. (Welf. & Inst. Code, § 6602, subd. (a).) The court is not required to make any findings on any of the other criteria of the Sexually Violent Predators Act at the probable cause hearing.” [5] Confinement cannot exceed two years unless a new petition is filed and an extended commitment is obtained from the court. (§ 6604.) The act also contains provisions for annual reviews and probable cause hearings to determine if the person’s mental health status has changed such that he or she is no longer a danger to others, and establishes the right of the State Department of Mental Health to seek the petitioner’s conditional or unconditional release. (§§ 6605, 6607, 6608.) [6] The statute was amended in 2000, and the prior language in subdivision (a)(1), stating that the previous qualifying conviction must have resulted in a determinate sentence, was removed. (§ 6600, as amended by Stats. 2000, ch. 643, § 1.) This amendment also added subdivision (a)(2), which delineates several other types of qualifying convictions in which the sentence is not necessarily determinate. (§ 6600, subd. (a)(2), as amended by Stats. 2000, ch. 643, § 1.) [7] The statute also specifies that “[i]f the victim of an underlying offense that is specified in subdivision (b) of Section 6600 is a child under the age of 14 and the offending act or acts involved substantial sexual conduct, the offense shall constitute a ‘sexually violent offense’ for purposes of Section 6600.” (§ 6600.1, subd. (a).) It defines “ ‘ [substantial sexual conduct’ ” as “penetration of the vagina or rectum of either the victim or the offender by the penis of the other or by any foreign object, oral copulation, or masturbation of either the victim or the offender.” (Id., subd. (b).) [8] T he SVPA does not provide any specific procedural requirements for the probable cause hearing. The parties do not dispute, however, that the hearing includes cross-examination of any experts relied on by the petitioner and the presentation of oral and written evidence bearing on probable cause. (In re Parker (1998) 60 Cal.App.4th 1453, 1468-1470 [71 Cal.Rptr.2d 167].) Although the petitioner is allowed, despite their hearsay nature, to present the contents of any reports that form the basis of the petition as evidence, the alleged sexual predator is allowed to cross-examine the expert concerning the evaluation and can call the expert to the stand for that purpose. (Id. at pp. 1469-1470.) The person named in the petition is thus allowed to “challenge the accuracy” of the evaluations by experts who found that he or she met the criteria for an SVP. (Id. at p. 1470.) The superior court may also permit the person named to call other witnesses who have relevant evidence to give bearing on the issue of probable cause. (Ibid) [9] At the trial stage, therefore, a judge or jury must find, beyond a reasonable doubt, that (1) the offender has been convicted of qualifying sexually violent offenses against at least two victims as defined in section 6600, subdivision (b); (2) the offender has a diagnosed mental disorder as defined in section 6600, subdivision (c); (3) the disorder makes it likely the offender would engage in sexually violent conduct if released; and (4) that this sexually violent conduct will be predatory in nature as defined in section 6600, subdivision (e). [10] We note that although it is necessary to find a mental disorder in order to prove likelihood of sexual violence in the future, the existence of a mental disorder alone is not sufficient to make such a finding. (See Ghilotti, supra, 27 Cal.4th at pp. 920-921 [“The requisite likelihood *249 of reoffense is thus a separate determination which does not inevitably flow from ... a predisposing mental disorder.” (Fn. omitted, italics added.)].) [11] We note that because this issue was not timely raised below, we may decline to consider this issue at this stage of the judicial process. (Cal. Rules of Court, rule 29(b)(1).) This issue was raised in Marentez’s petition for review, however, and the district attorney has responded substantively to the arguments in its briefs. As we recently noted in Ghilotti, this court has, on occasion, addressed issues that were not properly raised below, “where those issues were pure questions of law, not turning upon disputed facts, and were pertinent to a proper disposition of the cause or involved matters of particular public importance.” (Ghilotti, supra, 27 Cal.4th at p. 901, fn. 5.) Here the issue of the correct burden of proof at the SVPA hearing is a pure question of law and is of significant public importance in that it impacts the effective administration of this statutory scheme; we choose, therefore, to address the merits of Marentez’s arguments. [12] Evidence Code section 115 states in full: “ ‘Burden of proof means the obligation of a party to establish by evidence a requisite degree of belief concerning a fact in the mind of the trier of fact or the court. The burden of proof may require a party to raise a reasonable doubt concerning the existence or nonexistence of fact or that he establish the existence or nonexistence of fact by a preponderance of the evidence, by clear and convincing proof, or by *251 proof beyond a reasonable doubt. ffl] Except as otherwise provided by law, the burden of proof requires proof by a preponderance of evidence.” [13] We assume that Marentez’s argument, that these proceedings are “functional equivalents,” is analogous to the notion that persons utilizing these proceedings are “similarly situated” for purposes of equal protection analysis. [14] The issue of the meaning of “likely” at the trial stage of the S VP proceeding is before the court in People v. Roberge, review granted March 28, 2001, S094627. [15] Instead, Marentez unpersuasively argues that any error on the part of the trial court would not affect the outcome of the case, as the superior court’s determination was based on a lack of credibility of the prosecution’s witnesses, and not on their application of this term. Marentez fails to account for the fact that both the trial court and his own experts applied the wrong definition of “likely” to reach the conclusion that he did not meet the criteria for civil commitment under the SVPA. [16] In this case, the court did make an express finding that Marentez suffered from pedophilia that impaired his volitional capacity—a finding that we hold is necessary for a probable cause *260 determination. Marentez also did not dispute that his previous convictions were qualifying sexual offenses for purposes of the SVPA.