delivered the opinion of the Court.
Before the State can deprive a person of his freedom, either in a criminal trial or a civil commitment hearing, the State must satisfy *156a high standard of proof. Under the New Jersey Sexually Violent Predator Act (SVPA or Act), N.J.S.A. 30:4-27.24 to -27.38, a person previously convicted of a sexual offense can be civilly committed only if the State can establish by clear and convincing evidence that he suffers from a mental abnormality or personality disorder that makes him highly likely to commit a sexually violent offense. The experienced judges assigned to hear these cases have the difficult task of assessing expert testimony that often is in conflict, making factfindings about events described from varying viewpoints, and ultimately predicting the probability of a person’s future conduct. In the balance, an individual’s right to liberty is weighed against society’s interest in public safety.
In this case, R.F., when he was seventeen years old, engaged in sexual conduct with two children, ages twelve and thirteen. He pled guilty in adult court to endangering the welfare of both children and was sentenced to a five-year term at the Adult Diagnostic and Treatment Center at Avenel (ADTC or Avenel). Before R.F. completed his sentence, the State petitioned to have R.F. civilly committed under the SVPA.
At an SVPA commitment hearing, the Honorable Serena J. Perretti, J.S.C., sifted through a ream of documentary evidence and heard testimony from three expert witnesses, two for the State and one for R.F. Although Judge Perretti found that R.F. committed predicate sexual offenses and suffered from a personality disorder, she concluded that the State had not proven by clear and convincing evidence that R.F. was highly likely to engage in sexually violent behavior if not civilly committed. In coming to the decision that the State had not met the evidentiary standard for SVPA commitment, Judge Perretti made specific findings of fact and ultimately placed decisive weight on R.F.’s expert. She also made clear that R.F. is subject to parole supervision for life, N.J.S.A. 20:43-6.4; N.J.A.C. 10A:71-6.12. R.F., moreover, must also comply with a discharge plan prepared by the Special Treatment Unit (STU) where he has been civilly committed.
*157The Appellate Division reversed, determining that the opinions of the State’s experts were “well-supported by the record and amply substantiate the State’s petition for R.F.’s commitment under the SVPA.” Selecting the facts it deemed more credible, accepting the opinions it viewed more persuasive, and drawing its own inferences from the record, the panel came to a different conclusion than Judge Perretti.
The issue, however, is not whether members of the panel would have decided the case differently had they heard the case. Nor is the issue whether evidence in the record supports the opinions of the State’s experts. Rather, the issue is whether sufficient credible evidence in the record supports Judge Perretti’s findings. Those findings are entitled to deference, for Judge Perretti was not only intimately familiar with the case file but also had the unique opportunity to hear the witnesses, to judge their credibility, and to weigh their testimony—things that cannot be gleaned from the cold record.
Judge Perretti’s determination that the State did not establish by clear and convincing evidence that R.F. was highly likely to sexually reoffend unless institutionalized is supported by sufficient credible evidence in the record. Her findings are not clearly mistaken and are entitled to deference. We therefore reverse the Appellate Division and remand for proceedings consistent with this opinion.
In light of the passage of more than five years between Judge Perretti’s decision and our resolution today, we will stay the release of R.F. for thirty days to allow the State to file a new petition if there are any changed circumstances or conditions that would warrant civil commitment. On his release, R.F. will be subject to a discharge plan and parole supervision for life.
I.
R.F. was charged in two separate juvenile complaints with committing, in May and July 2004, first-degree aggravated sexual *158assaults against twelve-year-old A.M. and thirteen-year-old J.W. and other related offenses. R.F. was seventeen years old when the events that gave rise to the charges occurred. As part of a plea agreement, the case was waived to adult court where R.F. pled guilty to the lesser charges of third-degree endangering the welfare of A.M. and J.W., N.J.S.A. 2C:24-4(a).1 The State recommended that R.F. be sentenced to a five-year term at the ADTC based on an evaluation that he was a repetitive and compulsive offender.
In giving a factual basis to the charges, R.F. admitted that, on separate occasions, he had “sex” with A.M. and J.W. With each girl, he touched various parts of her body and placed his “penis like in her butt,” but “it fell out.” R.F., however, insisted that the sexual acts were consensual—a position he maintained in later interviews. During the plea colloquy, R.F. stated: “I just did it because they wanted to have sex ...[.] I didn’t want to do it, but I cannot say no to a girl for some stupid reason.” Nevertheless, he admitted that having sex with a minor was “wrong” and that he felt “bad.”2
That R.F. was cognitively impaired was evident at the time of the plea. Two of three psychiatrists who examined him declared that he was not competent to participate in the proceedings. Nevertheless, R.F. declared that he wished to proceed.
After that plea proceeding, at the correctional facility where he was incarcerated, R.F. threw a box at a corrections officer and then resisted several officers and spit on them. That incident led to R.F. pleading guilty to fourth-degree aggravated assault, N.J.S.A. 2C:12-1(b)(5).
*159In July 2005, R.F. was sentenced on the endangering and assault charges. An evaluation submitted by the ADTC to the court indicated that, although R.F. was deemed a repetitive and compulsive offender, he was considered amenable to treatment at the program at Avenel. R.F. did not object to an Avenel sentence. However, his attorney urged the court to find as a mitigating factor that R.F. “did not contemplate that his conduct would cause or threaten serious harm,” N.J.S.A. 2C:44-1(b)(2). The defense attorney stressed that R.F. was “barely competent to understand the proceedings” and that his conduct was the result of his psychiatric disorder, his “lack of education,” and “his inability to comprehend the consequences of [his] actions.”
Although the prosecutor commented on the seriousness of all three offenses, he acknowledged that R.F. had “a lot of mental limitations” and that R.F. was “in severe need of’ both sex offender and mental health treatment. The defense and prosecution agreed that the incident in the county jail was the product of “frustration” on R.F.’s part.
In imposing sentence, the court found, as aggravating factors, that R.F. had “acted out violently, or engaged in threats of violence,” and that the two young victims were “vulnerable” and “incapable of resisting his advances.” As mitigating factors the court found that R.F. was diagnosed with “mild mental retardation,” “deficits in his visual cognitive processes,” and “attention deficit, hyperactivity disorder” (ADHD), and that R.F. was classified as “neuro-psychiatrically compromised.” The court also noted R.F.’s youth and expression of remorse: “[H]e knew what he was doing was wrong, but he ... couldn’t resist the impulse to engage in the unlawful activity.”
The court sentenced R.F. to concurrent five-year terms on the endangering charges, to be served at the ADTC, and to a concurrent nine-month term on the assault charge, and imposed all requisite fines and penalties.3 The court also imposed parole *160supervision for life, N.J.S.A. 20:43-6.4, and noted that R.F. was subject to sex offender registration under N.J.S.A. 2C:7-2.
In May 2008, before R.F. completed his sentence, the State petitioned the Superior Court to have R.F. civilly committed under the SVPA. Attached to the State’s petition were the clinical certificates of two psychiatrists, each offering an opinion that R.F. “has serious difficulty controlling his sexually inappropriate impulses, and ... is highly likely to sexually reoffend.” Based on the petition and clinical certificates, a Superior Court judge found probable cause to believe that R.F. met the standards for commitment under the SVPA. R.F. was temporarily committed to the STU, a secure facility designated for sex offenders, pending an SVPA hearing.4
II.
Judge Perretti presided at the commitment hearing held in December 2008. During the hearing, a number of documents were presented to Judge Perretti, including statements that had been given by A.M. and J.W. to the police. Those statements described incidents of forcible sex, including the accusation by A.M. that R.F. grabbed her by the throat and the accusation by J.W. that R.F. threatened her with a knife.5 J.W.’s statement, however, indicated that R.F. and A.M. were in a consensual relationship with each other. The documentary record revealed that A.M. and J.W. were not strangers to R.F. but seemingly *161within his circle of friends—despite their significant age differences.
R.F. self-reported his sexual encounters with the two young girls. In late July 2004, R.F. told A.M.’s mother that he attempted to have sexual intercourse with her daughter, and in late August, he told J.W.’s brother that he had intercourse with J.W. Eventually, the mothers of the two girls reported the matters to the police. The varying accounts given by R.F., A.M., and J.W. about whether force was used are not reconcilable.
At the hearing, the State called two expert witnesses, Dr. Robert Harris, a psychiatrist, and Dr. Sean McCall, a psychologist. R.F. called one expert witness, Dr. Vivian Shnaidman, a psychiatrist.
A.
Dr. Harris testified that he interviewed R.F. three times, for approximately five to six hours in all, and found him cooperative but struggling, at times, to “maintain[ ] a line of thought.” Illustrative of this point is that, according to Dr. Harris, R.F. stated “that he was not attracted to younger kids,” “that he wasn’t attracted to girls over the age of 13,” “that some people are attracted to 17-year old girls and he doesn’t think it’s right and he would ... hurt those people,” and then, finally, that “his arousal to the 10 and 13 year old girls” is “not right.” Concerning R.F.’s mental capacity to participate in his earlier criminal proceedings, Dr. Harris was knowledgeable about the report of one psychiatrist who opined that R.F. was competent to participate but not about the reports of two other psychiatrists who opined that he was not.
Dr. Harris expressed his understanding that R.F. had “befriended” twelve-year-old A.M. and thirteen-year-old J.W. At the time, R.F. considered the girls to be his “peers.” R.F. reported to him that he became “sexually interested” in A.M. when she was nine or ten years old, which corresponded to when R.F. was fourteen or fifteen years old. From his review of the police reports and statements of A.M. and J.W. and his interviews with *162R.F., Dr. Harris hypothesized that R.F. was “targeting these two girls, creating an environment for them to play in ... and really going out of his way to groom [them].” Dr. Harris noted that the assault of A.M. occurred when R.F. entered the bathroom where she was concealing herself during a game of hide-and-seek with friends, including R.F. In an apparent reference to the assault on J.W., Dr. Harris stated that R.F. had “created this kind of clubhouse in an area where younger people would hang out.” (J.W.’s statement, however, indicated that she and her friends built the clubhouse.) As part of his grooming theory, Dr. Harris also emphasized that R.F. bought A.M. food and a bicycle— although the sources he relied on indicate that these gifts followed the incident in the bathroom.
Dr. Harris acknowledged that R.F. maintained that the sexual encounters with A.M. and J.W. were consensual. In his psychiatric analysis, Dr. Harris observed, on the one hand, that R.F.—who is in the “low average to borderline intellectual” range—“has this very reactive quality to him where ... he doesn’t think about what he’s doing. He acts in a way without any kind of planning.” On the other hand, Dr. Harris found that R.F.’s “sexual offenses were very well planned. He spent a great deal of time grooming and creating an environment to which he had access to these two girls.”
Dr. Harris diagnosed R.F. with pedophilia,6 ADHD,7 and antiso*163cial personality disorder.8 Dr. Harris concluded that R.F.’s pedophilia and antisocial personality placed him at a high risk to reoffend if he were not committed to the STU for sex offender treatment.
Because R.F. was “essentially 18 years old” at the time of the offenses, Dr. Harris did not believe that R.F.’s youth undermined the pedophilia diagnosis. He explained that R.F. had been attracted to A.M. for several years. Dr. Harris, however, did not elaborate on the point that R.F.’s attraction dated to a time when he was fourteen or fifteen years old. Instead, he cited to a statement by R.F. that he still had arousal to children. Dr. Harris stated that he did not know whether J.W. was prepubescent and that R.F. was unable “to figure out if she was [sexually] developed].” (A pedophilia diagnosis requires that the victim be prepubescent, according to the DSM-IV-TR, supra, at 571.)
Dr. Harris diagnosed R.F. with antisocial personality disorder because R.F. had a history of failing to control his behavior and because, he believed, the treatment at the ADTC was not successful. However, Dr. Harris agreed that R.F. had not been cited for any infractions since his commitment to the STU.
Dr. Harris measured R.F.’s risk of sexually reoffending with a diagnostic tool called the Static-99. R.F.’s scores of a four and a five placed him in the moderate- to high-risk category for sexually reoffending. Although Dr. Harris agreed that the Static-99 “should be used with caution” in grading juvenile offenders, he *164judged R.F. as an adult because he was just shy of eighteen years at the time of the offenses.9
B.
Dr. McCall, a psychologist at the STU, testified that, based on R.F.’s antisocial personality disorder, R.F. posed a high risk of sexually reoffending if not committed. Unlike Dr. Harris, Dr. McCall did not diagnose pedophilia because A.M. admitted to J.W. that she had a consensual relationship with R.F. He cited to the DSM-IV-TR, supra, at 572, which explicitly says not to diagnose pedophilia in “an individual in late adolescence involved in an ongoing sexual relationship with a 12- or 13-year-old.” Dr. McCall also could not substantiate paraphilia10 as a diagnosis because it was uncertain how “arousing” the events were to R.F. and because the activities did not occur over a period of six months.11
Dr. McCall settled on a diagnosis of antisocial personality disorder based on R.F.’s “fights in prison,” his “still breaking the rules,” his “being self-centered [and] kind of impulsive,” and his “willingness and interest in going after young girls.” In the breaking-the-rules category, Dr. McCall also pointed out that R.F. *165apparently had a consensual sexual encounter with another inmate at the ADTC. Dr. McCall emphasized that R.F. was on probation for a simple assault as a juvenile when he committed the endangering offenses. He believed that R.F.’s commission of those offenses while under probation supervision for simple assault “was a robust predictor of [sexual offense] recidivism.” Dr. McCall noted that R.F. tested in the “low average range of intellectual functioning” and that, given his Static-99 score of four, he fell within the group with a thirty-six percent likelihood of sexually reoffending within fifteen years.
Dr. McCall conceded that R.F. held firm that his sexual encounters with A.M. and J.W. were consensual; that he never admitted at any time to using force or a weapon; that neither A.M. nor J.W. alleged that R.F. anally penetrated them; that J.W. stated that she asked R.F. to build a fort in the woods (J.W.’s statement, however, indicates that she and her friends built the fort); that J.W., after having sex with R.F., believed she was having her period because of bleeding (thus suggesting that she was not prepubescent); that A.M. and J.W. admitted that they “hung out” with R.F.; and that R.F. saw himself as a “peer.” Dr. McCall also acknowledged that in R.F.’s statement to the police he admitted that it was wrong for him to have sex with A.M. and J.W. because of their age differences.
C.
Dr. Shnaidman, a psychiatrist who had been a consultant at the ADTC for more than five years, determined that R.F. did not fit the diagnostic profile for pedophilia or paraphilia. She ruled out those diagnoses because R.F. did not have “a specific arousal to prepubescent children” or nonconsensual sex. Dr. Shnaidman, however, diagnosed R.F. with conduct disorder.12 She also con*166ceded that he met the definition for antisocial personality disorder. She did not select the latter diagnosis because, although he committed the county-jail assault and other institutional infractions after he turned eighteen, R.F. did not mentally become eighteen on the day of his eighteenth birthday. Moreover, she did not weigh R.F.’s alleged consensual sexual encounter with another resident as an increased risk factor because the evidence did not suggest any “predatory or abusive” relationship. As she explained, “our standard for sex offenders somehow becomes much higher than the standards that we even hold ourselves to.”
Dr. Shnaidman also rejected the Static-99 as an effective diagnostic tool in assessing R.F.’s risk of sexually reoffending due to his cognitive limitations and the fact that the offenses were committed when he was a juvenile. The Static-99, in her estimation, is a test primarily for adults, and even then, is only “one piece of information.”
In Dr. Shnaidman’s opinion, R.F. posed a “fairly low” risk of sexually reoffending. She could not say the risk was “zero” but it was “not likely.” She came to that conclusion because, as a result of his criminal conviction and the treatment at the ADTC, R.F. knows that it is wrong to have sex with underage girls and to force someone to have sex.
In explaining the conduct that led to R.F.’s convictions, Dr. Shnaidman stressed R.F.’s cognitive limitations. She diagnosed him with “borderline intellectual functioning,” noting that at the age of twenty-two he had trouble reading The Cat in the Hat. She recounted that when she interviewed R.F., he did not “actually know” the charges to which he had pled guilty. She explained that R.F. saw himself as part of the victims’ peer group and considered his interactions to be consensual. According to Dr. Shnaidman, the evidence, including R.F.’s mental limitations, did not support the claim that R.F. was “grooming” A.M. and J.W.
In light of the differing accounts given by A.M., J.W., and R.F., Dr. Shnaidman acknowledged the difficulty in discerning the true circumstances of the encounters, but she emphasized that R.F. *167always denied using force and that A.M. told at least one friend (J.W.) that the sexual activity with R.F. was consensual.
In making her risk assessment, Dr. Shnaidman considered that R.F. would be subject to parole supervision for life. In her opinion, R.F. poses a low risk to reoffend and “seems to be motivated not to re-offend,” but nevertheless she stressed that providing structure and support for R.F. in the community would further reduce the risk of recidivism. Dr. Shnaidman expressed that anyone “involved in sex offender treatment should be transitioned into the community with some kind of supervision and some kind of services,” and in R.F.’s case vocational training and counseling would serve as prevention therapy.
III.
Before rendering her decision, Judge Perretti not only heard the testimony of the three expert witnesses, but also reviewed the relevant documentary evidence, such as the victims’ and R.F.’s statements, R.F.’s plea and sentencing transcripts, and the expert reports. She viewed the evidence through the prism of the governing legal principles, keeping in mind that the State had the burden of establishing by clear and convincing evidence the elements required for SVPA commitment.
Judge Perretti first concluded that R.F. pled guilty to predicate offenses that made him eligible for commitment under the SVPA. She noted that the record revealed that R.F. had penetrated thirteen-year-old J.W. and had sexual contact with twelve-year-old A.M. Based on these circumstances, Judge Perretti concluded that R.F.’s convictions for third-degree endangering the welfare of those children were sexually violent offenses under the Act.
Second, Judge Perretti concluded that, based on his conduct in the community and in custody as well as the experts’ testimony, R.F. had juvenile conduct disorder. She determined that Dr. Harris’s diagnosis of pedophilia did “not squarely fit the criteria of the DSM IV.” In that regard, Judge Perretti observed that “there may have been an ongoing relationship between [R.F.] and A.M.” *168and that it is not known whether she was prepubescent. She also noted that both Dr. McCall and Dr. Shnaidman rejected the diagnosis of pedophilia. Judge Perretti, moreover, did not find paraphilia as an appropriate diagnosis based on Dr. Shnaidman’s testimony. In support of the juvenile conduct disorder diagnosis, Judge Perretti mentioned that R.F. had been “assigned to the mental health program at [the] ADTC and reportedly had bouts of anger which resulted in his getting into trouble.” Referencing his therapist’s report, she stated that R.F. had “demonstrated distorted thinking and had only a limited understanding of his crime”; that he “had deficits in intellectual functioning, social skills, communication skills and developing relationships”; that he “had poor impulse control which resulted in [his] acting out behaviors within the [ADTC]”; and that he “had made only limited progress in treatment.” She also observed that R.F. received a number of disciplinary citations at the ADTC, including an incident in which R.F. was found in a portable bathroom with another inmate “under suspicious circumstances.”
Despite the predicate-offense and personality-disorder findings, Judge Perretti held that the State did not establish by clear and convincing evidence that R.F. was “highly likely to commit a sexually violent offense.” In reaching this conclusion, Judge Perretti placed great weight on Dr. Shnaidman’s opinion that R.F.’s risk to reoffend was “ ‘fairly low1 ” and that R.F. had learned that it is wrong to have sex with someone under age. She also gave weight to the fact that R.F. disclosed his wrongdoing to the victims’ families: “It would seem that [R.F.] was conscious at the time that he made his disclosures of the wrongness of his acts” and that the disclosures “indicate[ ] a resolution to desist.” Further, Judge Perretti highlighted that R.F. would be subject to parole supervision for life, “which affords some protection to the public and may act as a deterrent.”
Judge Perretti also observed that some of the testimony of the State’s experts “appears [to be] based upon an exaggeration or misunderstanding of the circumstances which led to their eonelu-*169sions.” For example, Dr. Harris diagnosed R.F. with pedophilia partly based on the assumption that R.F. had engaged in grooming behavior. Yet, as Judge Perretti pointed out, the documentary evidence revealed that R.F. had not—as believed by Dr. Harris—built the fort that Dr. Harris saw as a way of luring children. Judge Perretti, moreover, noted that the “gifts” from R.F. to A.M. “followed the sex incident”13 and supposedly were “offered and received to keep her quiet.” To her mind, the State did not demonstrate that A.M. or J.W. had “any ongoing fear of [R.F.] who continued to be part of the circle of friends.”
Judge Perretti further explained that Dr. Harris’s opinions “depend[ed] greatly upon [R.F.’s] self-disclosures” during his interviews with R.F. However, some of R.F.’s statements referenced by Dr. Harris, in the judge’s view, were “incomprehensible,” “babbling,” and showed “a seeming confusion as to time and persons.”
In addition, Judge Perretti maintained that Dr. McCall’s initial paraphilia diagnosis was based on an unfounded assumption. Dr. McCall assumed that J.W. sustained injuries during her sexual encounter with R.F. based on the presence of bloodstains. J.W., however, stated to the police that she believed the blood was a result of “her monthly period.”
Judge Perretti also noted that the prosecutor’s office did not pursue aggravated sexual assault charges but rather allowed R.F. to plead guilty to third-degree endangering charges with a five-year maximum exposure. Yet, the State’s petition for civil commitment, if granted, “in this case is tantamount to life in custody.” That raised in her mind issues of proportionality.
Judge Perretti found both Dr. Harris and Dr. Shnaidman “equally well-qualified” and that the difference of opinions between these “highly respected” experts was “itself a matter generating doubt.” To Judge Perretti, there was a distinction between *170a reasonable prediction that R.F. would “get in trouble if released now into the community”—a proposition she accepted—and a finding by clear and convincing evidence that R.F. was “highly likely to commit a sexually violent offense in the foreseeable future”—a proposition she did not accept.
Last, Judge Perretti had little doubt that R.F. would “require many social services if he is to peacefully negotiate life in the community.” She maintained, however, that the “State must step up to the plate now and cannot simply hide [R.F.] in the Special Treatment Unit.” Judge Perretti dismissed the State’s SVPA petition but stayed her order pending appeal.
IV.
In an unpublished per curiam opinion, the Appellate Division reversed the dismissal of the State’s petition, thereby directing that R.F. be civilly committed. The panel concluded “that the record does not support the trial court’s determination that R.F. does not qualify for [SVPA] commitment.” Although the panel acknowledged that “[t]he scope of appellate review of the trial court’s findings is extremely narrow,” citing In re Civil Commitment of V.A., 357 N.J.Super. 55, 63, 813 A.2d 1252 (App.Div.), certif. denied, 177 N.J. 490, 828 A.2d 917 (2003), the panel made its own findings of fact and rejected those made by Judge Perretti.
The panel deemed it error for the trial court to consider that R.F.—with his limited cognitive ability—viewed himself as the girls’ peer. It found that'R.F.’s “behavior was calculating and predatory” and that “[h]e sought out the girls with the intention of engaging in sexual activity, with or without their assent.” Thus, accepting in full the victims’ accounts, it determined that R.F. “violently sexually assaulted” A.M. in a locked bathroom and lured J.W. to “a secluded area [and] then forcibly sexually assaulted her at knife point.” The panel maintained that R.F.’s “denial or minimization of the harm he caused his victims ... revealfed] his *171distorted and pathological perception that he was acting in an age-appropriate manner when he assaulted the girls.”
The panel faulted the trial court for “unduly discounting] the testimony of the State’s expert witnesses,” stressing that their opinions were “well-supported by the record.” It was dismissive of Dr. Shnaidman’s opinion for failing to reckon “the danger posed to the community at large by the precipitous release” of R.F. Last, the panel determined that, based on his history of noncompliance with treatment, R.F. was “highly likely to re-offend if released under the conditions suggested by Dr. Shnaidman.”
We granted R.F.’s petition for certification. In re Civil Commitment of RF., 212 N.J. 288, 53 A.3d 663 (2012).
V.
A.
R.F. submits that the Appellate Division failed to adhere to the limited scope of its review when it disregarded the trial court’s findings. According to R.F., the trial court heard extensive testimony from three different experts, considered voluminous documentary materials, and then reached a decision fully supported by the evidence. R.F. argues that the trial court did not have to credit the ultimate opinions of the State’s experts or to reject the opinion offered by Dr. Shnaidman. Rather, the court had to exercise its own independent judgment in making findings of fact and determining whether R.F. was highly likely to sexually reoffend. The panel, R.F. maintains, did not give Judge Perretti’s findings the utmost deference and merely substituted its own findings. Last, R.F. believes that the SVPA allows the court to place conditions on R.F.’s release from the STU or to make recommendations in regard to parole supervision for life.
B.
The State urges this Court to affirm the Appellate Division, reasoning that the trial court’s determination was a “manifestly *172mistaken exercise of discretion.” The State argues that Judge Perretti “abused her discretion when she rejected the diagnosis of pedophilia” because “there was ample evidence in the record to support a diagnosis of pedophilia.” It submits that Judge Perretti speculated in suggesting that R.F.’s disclosures to Dr. Harris may have been unreliable and that J.W. may not have been prepubescent.
The State also declares that the evidence “overwhelmingly established that R.F. is ‘likely to engage in acts of sexual violence if not confined’ ” to the STU, quoting N.J.S.A. 30:4-27.26. The State questioned the trial court’s finding that R.F. could be released subject to parole supervision for life without endangering the public, given that he had “sexually assaulted the victims while on supervised probation.”
Last, the State professes that this Court has no jurisdiction to place conditions on R.F.’s discharge in the event that it reinstates Judge Perretti’s decision. Based on its reading of N.J.S.A. 30:4-27.32(b), the State suggests that court-imposed conditions can only be implemented and enforced if R.F. is committed under the SVPA and later found by the commitment court no longer to be a sexually violent predator.
VI.
The preeminent issue in this case concerns the scope of appellate review in commitment cases involving the Sexually Violent Predator Act, N.J.S.A. 30:4-27.24 to -27.38. Trial judges who handle SVPA commitment hearings generally possess expertise and experience in highly complex matters where credibility decisions must be made, expert psychiatric testimony assessed, future conduct predicted, and individual liberty weighed against public safety. The level of deference that is accorded to trial court decisions in SVPA cases is at the heart of the conflict between R.F. and the State. Resolving that issue requires an understanding of the SVPA, to which we turn now.
*173A.
The SVPA permits the State to involuntarily commit “a person who has been convicted ... of a sexually violent offense” who “suffers from a mental abnormality or personality disorder that makes the person likely to engage in acts of sexual violence if not confined in a secure facility for control, care and treatment.” N.J.S.A 30:4-27.26. At the commitment hearing, the State must establish three elements: (1) that the individual has been convicted of a sexually violent offense, ibid.; (2) that he suffers from a mental abnormality or personality disorder, ibid.; and (3) that as a result of his psychiatric abnormality or disorder, “it is highly likely that the individual will not control his or her sexually violent behavior and will reoffend,” In re Commitment of W.Z., 173 N.J. 109, 130, 801 A.2d 205 (2002). Although the first two elements derive directly from the statute, to comport with substantive due process concerns, this Court interpreted the third statutory element as requiring the State to show that a person is “highly likely,” not just “likely,” to sexually reoffend. Ibid.
The State bears the burden of proving all three elements by clear and convincing evidence. N.J.S.A 30:4-27.32(a). Clear and convincing evidence is evidence that produces “a firm belief or conviction” that the allegations are true; it is evidence that is “so clear, direct and weighty and convincing” that the factfinder can “come to a clear conviction” of the truth without hesitancy. In re Jobes, 108 N.J. 394, 407, 529 A.2d 434 (1987) (quoting State v. Hodge, 95 N.J. 369, 376, 471 A.2d 389 (1984)). The terms of the statute must be strictly met. The State cannot confine a person because it is reasonably likely that he will not be able to abide by all of society’s laws or norms. SVPA commitment is limited to those who are highly likely to sexually reoffend. Cf. Kansas v. Crane, 534 U.S. 407, 412, 122 S.Ct. 867, 870, 151 L.Ed.2d 856, 862 (2002) (noting “the constitutional importance of distinguishing a dangerous sexual offender subject to civil commitment ‘from other dangerous persons who are perhaps more properly dealt with exclusively through criminal proceedings’” (quoting Kansas v. *174Hendricks, 521 U.S. 346, 360, 117 S.Ct. 2072, 2081, 138 L.Ed.2d 501, 514 (1997))).
At issue in this case is not whether R.F. committed a predicate sexual offense under the SVPA or even whether he suffers from a personality disorder or mental abnormality—although the nature of the disorder or abnormality is contested. The key dispute is whether based on the sexual offenses committed, the disorder or abnormality, and his juvenile history and institutional infractions, R.F. is highly likely to commit another sexual offense.
We now turn to the standard governing appellate review.
B.
“The scope of appellate review of a commitment determination is extremely narrow.” In re D.C., 146 N.J. 31, 58, 679 A.2d 634 (1996) (citing State v. Fields, 77 N.J. 282, 311, 390 A.2d 574 (1978)). The judges who hear SVPA eases generally are “specialists” and “their expertise in the subject” is entitled to “special deference.” See In re Civil Commitment of T.J.N., 390 N.J.Super. 218, 226, 915 A.2d 53 (App.Div.2007). The final decision whether a person previously convicted of a sexually violent offense is highly likely to sexually reoffend “lies with the courts, not the expertise of psychiatrists and psychologists. Courts must balance society’s interest in protection from harmful conduct against the individual’s interest in personal liberty and autonomy.” See D.C., supra, 146 N.J. at 59, 679 A.2d 634 (stating principles that apply in ordinary civil commitment hearings). A trial judge is “not required to accept all or any part of [an] expert opinion[ ].” Id. at 61, 679 A.2d 634. The ultimate determination is “a legal one, not a medical one, even though it is guided by medical expert testimony.” Id. at 59, 679 A.2d 634.
We give deference to the findings of our trial judges because they have the “opportunity to hear and see the witnesses and to have the ‘feel’ of the ease, which a reviewing court cannot enjoy.” State v. Johnson, 42 N.J. 146, 161, 199 A.2d 809 (1964). *175An appellate court should not overturn a trial court’s findings because it “might have reached a different conclusion were it the trial tribunal” or because “the trial court decided all evidence or inference conflicts in favor of one side” in a close case. Id. at 162, 199 A.2d 809.
Accordingly, an appellate court should not modify a trial court’s determination either to commit or release an individual unless “the record reveals a clear mistake.” D.C., supra, 146 N.J. at 58, 679 A.2d 634 (citing Fields, supra, 77 N.J. at 311, 390 A.2d 574); see Johnson, supra, 42 N.J. at 162, 199 A.2d 809 (stating that trial court’s findings should be disturbed only if so clearly mistaken “that the interests of justice demand intervention” and only then should appellate court “appraise the record as if it were deciding the matter at inception and make its own findings and conclusions”). So long as the trial court’s findings are supported by “sufficient credible evidence present in the record,” those findings should not be disturbed. Johnson, supra, 42 N.J. at 162, 199 A.2d 809; see In re Civil Commitment of J.M.B., 197 N.J. 563, 597, 964 A.2d 752 (stating that appellate courts must defer where there is “substantial, credible evidence to support the court’s findings”), cert, denied, 558 U.S. 999, 130 S.Ct. 509, 175 L.Ed.2d 361 (2009).
VII.
In light of these governing standards, we now must determine whether the Appellate Division had a proper basis to overthrow the findings of the trial court. In the end, we conclude that the panel overstepped the narrow scope of appellate review applicable in this case. The major flaw in the panel’s analysis is that instead of surveying the record to see whether there was sufficient credible evidence to support Judge Perretti’s factfindings, the panel assembled bits and pieces of the record that supported the State’s case. The issue was not, as the panel stated, whether the opinions of the State’s experts were “well-supported by the record” and thus “amply substantiate^] the State’s petition.” That *176the Appellate Division would have come to a different conclusion had it sat as the trier of fact is not a basis for overturning the trial court’s decision.
The State bore the burden of proving by clear and convincing evidence each of the elements for commitment under the SVPA. Judge Perretti found that the State satisfied the first two elements. She found, however, that the State did not present clear and convincing evidence of the third and decisive element—that R.F. was highly likely to sexually reoffend. In reaching that conclusion, Judge Perretti stated her reasons for rejecting the ultimate opinions of the State’s two experts and accepting the opinion of Dr. Shnaidman. Sufficient credible evidence is present in the record to support Judge Perretti’s decision.
A.
Judge Perretti found that R.F.’s plea of guilty to endangering the welfare of twelve-year-old A.M. and thirteen-year-old J.W. in 2004, when R.F. was seventeen years old, constituted prior “sexually violent offenses” under the catch-all provision of the SVPA, N.J.S.A 30:4-27.26(b).14 No one disputes that finding.
Judge Perretti also found that R.F. suffered from a personality disorder—the second element necessary for commitment under the SVPA. That finding is not contested. Nevertheless, the State argues that Judge Perretti abused her discretion in rejecting Dr. *177Harris’s pedophilia diagnosis. Yet, the State’s own expert, Dr. McCall, as well as Dr. Shnaidman, did not believe that pedophilia was substantiated—and, according to Judge Perretti, a pedophilia diagnosis did “not squarely fit the criteria of the DSM IV.” Moreover, we cannot say that Judge Perretti clearly erred by accepting the opinion of Dr. Shnaidman who found that a paraphilia diagnosis was inappropriate because R.F. did not have “a specific arousal to prepubescent children.” Judge Perretti was not bound to adopt the State’s opinions; she was required to exercise her independent judgment in making findings that, as here, were supported by the record. See D.C., supra, 146 N.J. at 59,679 A.2d 634.
B.
The critical finding of Judge Perretti was that the State failed to show by clear and convincing evidence that R.F., if released from confinement, was highly likely to commit another violent sexual offense. In coming to that conclusion, Judge Perretti took into account R.F.’s youth, the cognitive limitations that led him to perceive himself as a peer of twelve- and thirteen-year old girls, and his self-reporting of his sexual encounters to the mother of A.M. and the brother of J.W. She also placed great weight on Dr. Shnaidman’s opinion that R.F. had learned that having sexual relations with someone under age is wrong and that his risk of sexually reoffending was “fairly low.”
Judge Perretti understood that many facts were in dispute—a point not fully grasped by the Appellate Division. The experts disagreed about whether the evidence established that the girls were prepubescent, an important factor in assessing the nature of R.F.’s disorder; whether, given R.F.’s cognitive limitations and age, the Static-99 was an appropriate diagnostic tool for measuring his risk of sexually reoffending; and whether R.F.’s viewing the girls as his peers increased or decreased the risk that he would sexually reoffend. Additionally, there were conflicting accounts about whether violence was used during the sexual encoun*178ters and misunderstandings about whether R.F. built a fort and then used it to lure children. Before the SVPA commitment hearing, no trial ever resolved these disputed facts and issues. That difficult task was left to Judge Perretti. She sifted through the documentary evidence, heard the testimony of the experts, and came to her factual findings and legal conclusions.
Judge Perretti had a full understanding of the factual limitations in the record, and that led her to have doubt about whether the State had carried its burden. In its opinion, the Appellate Division exceeded its scope of review because it did not canvass the record for credible evidence to support Judge Perretti’s factfindings. Instead, it drew its own inferences from the record and made its own factfindings—different from those of Judge Perretti—that R.F. acted in a “predatory” manner, violently sexually assaulted the girls, and minimized the harm he caused the victims. However, a mere disagreement with the trial court’s factfindings cannot be the basis for substituted factfindings by an appellate court. Judge Perretti, moreover, was authorized to discount an expert opinion that she believed was at odds with the record and not as well grounded as another expert opinion.
In this case, a judge who sat regularly on SVPA cases and who was a specialist in the area came to a reasoned conclusion based on sufficient credible evidence in the record. She judged a cognitively impaired young man who was convicted of having sexual relations with underage minors and who had a prior juvenile record and a history of institutional infractions. Judge Perretti did not look at the case through rose-colored glasses. She knew that the raw truth sometimes is not easily discernible. She did what we expect of judges—she viewed difficult facts against the applicable law.
Judge Perretti understood that a reasonable prediction could be made that, given his disorder and background, without help, R.F. would “get in trouble if released now into the community.” But the SVPA only permits for the civil commitment of those who are highly likely to commit a sexually violent offense if released. See *179W.Z., supra, 173 N.J. at 129-30, 801 A.2d 205; cf. Crane, supra, 534 U.S. at 412, 122 S.Ct at 870, 151 L.Ed.2d at 862. That she could not find by clear and convincing evidence.
Judge Perretti also considered that R.F. is subject to a multiplicity of conditions and restrictions through parole supervision for life, which will minimize any potential threat to public safety. See N.J.S.A. 20:43-6.4; N.J.A.C. 10A:71-6.12(b). The sweeping supervision to which R.F. will be subject after his release requires R.F. to: live in an approved residence, N.J.A.C. 10A:71-6.12(d)(5)-(6); receive permission to leave the state, N.J.A.C. 10A:71-6.12(d)(7); refrain from possessing weapons, N.J.A.C. 10A:71-6.12(d)(8)-(9), and from possessing or using controlled dangerous substances, N.J.A.C. 10A:71-6.12(d)(10); submit to drug and alcohol testing, N.J.A.C. 10A:71-6.12(d)(13), and psychological testing, N.J.A.C. 10A:71-6.12(d)(11); complete appropriate counseling or treatment, N.J.A.C. 10A:71-6.12(d)(12), including any therapy specified by the staff at the ADTC, N.J.A.C. 10A:71-6.12(g); obtain permission before accepting employment, N.J.A.C. 10A:71-6.12(d)(14); notify his parole officer if he becomes unemployed, N.J.A.C. 10A:71-6.12(d)(15); avoid any contact with A.M. or J.W., N.J.A.C. 10A:71-6.12(d)(16); comply with any curfew, N.J.A.C. 10A:71-6.12(d)(17); and submit to warrantless searches by his parole officer, N.J.A.C. 10A:71-6.12(d)(21).
The terms of R.F.’s supervised release also require that he generally not contact or attempt to contact a minor, N.JAC. 10A:71-6.12(e)(1)-(2), and not live with a minor without permission, N.JAC. 10A:71-6.12(e)(3).
What is more, “special conditions” can be imposed on R.F. if “such conditions would reduce the likelihood of recurrence of criminal behavior.” N.J.A.C. 10A:71-6.12(n). These supervision requirements are also accompanied by mandatory registration requirements under Megan’s Law. N.J.S.A. 2C:7-2. It was appropriate for Judge Perretti to consider these conditions in reaching her decision.
*180VIII.
A.
Judge Perretti acknowledged that R.F., who at the time had been confined for over four years, first at the county jail, then at the ADTC, and later at the STU, would “require many social services if he is to peacefully negotiate life in the community.” That, in the free world, an individual may need assistance in housing, in vocational training, in mental health counseling, and in other life skills is not a reason for his continued commitment in the STU. That is what Judge Perretti meant when she stated: “The State must step up to the plate now and cannot simply hide [R.F.] in the Special Treatment Unit.”
Upholding Judge Perretti’s decision means that R.F. must be released with an appropriate discharge plan prepared by the STU staff. See N.J.S.A. 30:4-27.37 (“A person discharged by the court shall have a discharge plan prepared by the treatment team at the facility designated for the custody, care and treatment of sexually violent predators____”); N.J.S.A. 30:4-27.32(b) (stating that person found not to be sexually violent offender must be released with “discharge plan”). That discharge plan, crafted by staff of the STU, presumably will provide for the services and counseling necessary for R.F.’s successful reintegration into the community.
The Department of Human Services, which operates the STU, must decide whether to provide the services necessary for former committees, such as R.F., to live successfully in the free world. Surely, the Department will want to maximize the likelihood of R.F.’s reintegration into the community and minimize the risk of recidivism. Our civil commitment jurisprudence has emphasized the importance of “provid[ing] the needed level of care in the least restrictive manner,” In re S.L., 94 N.J. 128, 141, 462 A.2d 1252 (1983), and not infringing on an individual’s “liberty or autonomy any more than appears reasonably necessary to accomplish” the State’s goals of public safety and effective treatment, State v. Krol, 68 N.J. 236, 261-62, 344 A.2d 289 (1975).
*181Significantly, the SVPA provides for a “conditional discharge” of a committee when the Department of Human Services so recommends and “the court finds that the person will not be likely to engage in acts of sexual violence because the person is amenable to and highly likely to comply with a plan to facilitate the person’s adjustment and reintegration into the community.” N.J.S.A. 30:4-27.32(c)(1). The goal of a discharge plan for an individual who is conditionally discharged is to “render involuntary commitment as a sexually violent predator unnecessary for that person.” Ibid. Although R.F.’s release follows from a different provision of the SVPA, the goal of his discharge plan prepared by the STU should also be to facilitate his “adjustment and reintegration into the community.” See ibid.
R.F. has lived over nine years—his entire adult life—in the custody of the State. The record is clear that R.F. is a cognitively impaired young man who requires assistance when he is released into the community. Clearly, the STU staff and the Department of Human Services will be in the best position to decide what services and counseling R.F. will require to successfully navigate outside the confines of a State institution. Accordingly, we remand to the civil commitment court to allow the “treatment team” at the STU to formulate an appropriate discharge plan.
B.
Our difference with the dissent is that it pays lip service to the standard of review by ignoring evidence and testimony that supports the findings of Judge Perretti and by cherry-picking facts that would support the State’s petition. In addition, the dissent has a mistaken understanding of some portions of the record. For example, the dissent does not take into account that Dr. Shnaid-man found and Judge Perretti accepted that R.F.—having served his prison term—understands that it is wrong to have sexual relations with underage individuals.
The dissent also misapprehends our citation to the subsection of the SVPA that references a “conditional discharge,” N.J.S.A. 30:4-*18227.32(c)(1). We know that R.F. is not a candidate for conditional discharge under that statute. Our only point is that whether a discharge plan is crafted under N.J.S.A. 30:4-27.32(c)(1) or under N.J.S.A. 30:4-27.32(b), when the court finds a person, such as R.F., should be released, the goal is “to facilitate the person’s adjustment and reintegration into the community so as to render involuntary commitment” unnecessary, N.J.S.A. 30:4-27.32(c)(1).
The complex and difficult judgment calls to be made after hearing the testimony of the experts and sifting through the evidence was for the trial judge, and we must not second-guess those calls unless they are clearly mistaken and unsupported by the evidence. It is here where we part ways with the dissent.
C.
R.F. has been detained at the STU for over five years without any judicial review of his mental or behavioral status. This Court has not been informed of R.F.’s current status or of his progress at the STU over the lengthy history of this appeal. We cannot foreclose the possibility that circumstances or conditions that might have a bearing on whether R.F. is highly likely to sexually reoffend have changed since Judge Perretti’s ruling. Our decision does not deprive the State of the right to re-petition for SVPA commitment based on changed circumstances and conditions.15 Nevertheless, if there is a basis to re-petition, it may not be used as an occasion to re-litigate or collaterally attack Judge Perretti’s findings. Those findings are not to be revisited.
In light of the unusual posture of this appeal, we will stay the discharge of R.F. for thirty days. During that period, a discharge plan shall be prepared and the State can decide whether there is any ground to re-petition based on changed circumstances and conditions since Judge Perretti’s decision. If the State chooses to re-petition, the civil commitment court shall provide R.F. with a *183prompt probable-cause determination, see N.J.S.A. 30:4-27.28(f), and hearing, see N.J.S.A. 30:4-27.29(a) (requiring hearing “within 20 days from the date of the temporary commitment order”).
IX.
For the reasons expressed, we reverse the judgment of the Appellate Division and reinstate the decision rendered by the trial court, subject to the modifications set forth in this opinion. We remand to the trial court for proceedings consistent with this opinion.
R.F. waived his right to have the charges presented to a grand jury and pled guilty to the endangering charges as presented in an accusation. In accordance with the plea bargain, the State dismissed the juvenile charges.
In statements given to the police, both A.M. and J.W. claimed that their sexual encounters with R.F. were not consensual.
R.F. was given jail credit for 310 days served in the county jail.
The Attorney General is empowered to initiate proceedings to involuntarily commit persons who have been identified as sexually violent predators. N.J.S.A. 30:4-27.28. If there is probable cause to believe that a person is a sexually violent predator, a court orders an initial commitment until a trial-like hearing is conducted. N.J.S.A. 30:4-27.28(g). At the plenary hearing, the State must demonstrate "by clear and convincing evidence that the person needs continued involuntary commitment as a sexually violent predator.” N.J.S.A. 30:4-27.32.
These statements were made part of the presentence report used by the trial court in fixing R.F.’s sentence.
The Diagnostic and Statistical Manual of Mental Disorders characterizes pedophilia in the following way:
The paraphilic focus of Pedophilia involves sexual activity with a prepubescent child (generally age 13 years or younger). The individual with Pedophilia must be age 16 years or older and at least 5 years older than the child. For individuals in late adolescence with Pedophilia, no precise age difference is specified, and clinical judgment must be used; both the sexual maturity of the child and the age difference must be taken into account.
[American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders 571 (4th ed. Text Revision 2000) [hereinafter DSM-IV-7»].]
*163The latest edition of the DSM was published in 2013. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders (5th ed. 2013).
"The essential feature of [ADHD] is a persistent pattern of inattention and/or hyperactivity-impulsivity that is more frequently displayed and more severe than is typically observed in individuals at a comparable level of development.” DSM-IV-TR, supra, at 85.
Antisocial personality disorder is defined by "a pervasive pattern of disregard for, and violation of, the rights of others that begins in childhood or early adolescence and continues into adulthood.” DSM-IV-TR, supra, at 701.
The Static-99 is an actuarial test used to estimate the probability of sexually violent recidivism in adult males previously convicted of sexually violent offenses. See Andrew Harris et al., Static-99 Coding Rules Revised-2003 5 (2003). This Court has explained that actuarial information, including the Static-99, is "simply a factor to consider, weigh, or even reject, when engaging in the necessary factfinding under the SVPA.” In re Commitment of R.S., 173 N.J. 134, 137, 801 A.2d 219 (2002).
The DSM-IV-TR describes paraphilia as "recurrent, intense sexually arousing fantasies, sexual urges, or behaviors generally involving 1) nonhuman objects, 2) the suffering or humiliation of oneself or one's partner, or 3) children or other nonconsenting persons that occur over a period of at least 6 months.” DSM-IV-TR, supra, at 566.
Dr. McCall initially believed that R.F. suffered from pedophilia and paraphilia, but ultimately abandoned those diagnoses on further consideration.
"The essential feature of Conduct Disorder is a repetitive and persistent pattern of behavior in which the basic rights of others or major age-appropriate societal norms or rules are violated.” DSM-IV-TR, supra, at 93.
Judge Perretti mistakenly referred to J.W. here instead of A.M.
N.J.S.A. 30:4-27.26(a) specifically designates such offenses as aggravated sexual assault and sexual assault as sexually violent offenses. The offenses classified in subsection (a), however, are not an exhaustive list. N.J.S.A. 30:4-27.26(b) expands sexually violent offenses to include "any offense for which the court malees a specific finding on the record that, based on the circumstances of the case, the person’s offense should be considered a sexually violent offense.” A finding of a sexually violent offense under subsection (b) "requires substantially equivalent conduct to the conduct captured by the offenses listed in subsection (a).” J.M.B., supra, 197 N.J. at 595, 964 A.2d 752. Here, although R.F. pled guilty to endangering the welfare of a child, the substantially equivalent conduct would be sexual assault against J.W., N.J.S.A. 2C: 14—2(c)(4), and sexual assault against A.M., N.J.S.A. 2C:14-2(b).
N.J.S.A. 30:4-27.28 describes the procedures by which the Attorney General may initiate involuntary commitment proceedings under the SVPA.