¶1 At a trial to determine whether Charles Post was a sexually violent predator (SVP), the State introduced evidence about the treatment that would be available to Post if he were civilly committed and the possibility of future release to a less restrictive alternative. The jury found that Post was an SVP, and Post appealed. Post argues that evidence about the treatment and possibility of future release if he is committed as an SVP is irrelevant to the determination of whether he is currently an SVP. We agree and find that the erroneous admission of this evidence was not harmless. Additionally, we hold that evidence that a person who commits a “ ‘[rjecent overt act,’ ” as defined by RCW 71.09.020(12), could be subject to a new SVP commitment petition is relevant in an SVP determination trial. We therefore affirm the judgment of the Court of Appeals and remand the case for further proceedings.
*306FACTS
¶2 On January 13, 2003, one day before Post was scheduled to be released from confinement for his 1987 first degree burglary and first degree rape convictions, the King County prosecutor filed a petition alleging that Post was an SVP. On March 7, 2003, the superior court found probable cause to believe that Post was an SVP and ordered his continued detention at the Special Commitment Center (SCC) pending trial. While at the SCC, Post voluntarily participated in the first two phases of sex offender treatment therapy. Prior to the State’s filing its petition, Post had also voluntarily participated in the Twin Rivers sex offender treatment program at the Monroe Correctional Complex. From February 2001 through early June 2001, Post participated in group therapy. In June 2001, Post was removed from group therapy but continued participating in individual therapy until he was removed from treatment based on an unrelated disciplinary infraction.
¶3 Post went through two SVP determination trials. The first trial began on June 30, 2004, and ended in a mistrial on July 14, 2004, when the jury was unable to reach a verdict. At this first trial, the State made no reference to treatment available to committed SVPs at the SCC. Post’s second SVP determination trial began on November 15, 2004, and ended on December 16, 2004, with a jury verdict finding that the State had proved beyond a reasonable doubt that Post was an SVP. This trial was closely contested. The State called 12 witnesses, including a psychologist who had interviewed Post and 3 persons who had worked with Post as part of his sex offender treatment therapy. Post, in turn, called 23 witnesses, including multiple experts. A major component of Post’s defense was that he had a voluntary treatment plan that would reduce the likelihood that he would commit another act of sexual violence if released unconditionally into the community.
¶4 In a departure from its approach during the first trial, the State, in the second trial, sought to introduce evidence *307of the treatment that would be available to Post at the SCC if he were committed. Post objected to this evidence, raising motions in limine to exclude all such evidence and also to exclude all evidence that Post might be released to a less restrictive alternative in the future. The trial court denied both motions. Post also proposed limiting instructions on the use of such testimony, but the trial court denied these instructions. In addition, the trial court prohibited Post from introducing evidence that, if released into the community, he could be subject to a new SVP commitment petition if he were to commit a recent overt act.
¶5 Testimony about treatment available at the SCC came largely from two sources. James Anderson, a forensic therapist at the SCC, testified at length about the six phases of treatment offered at the SCC and the courses required in each phase. Anderson also testified about the conditions placed on persons conditionally released from the SCC. The jury submitted numerous questions to Anderson, asking, among other things, how long the six-phase treatment program had been in use at the SCC, how many SVPs had been committed to and subsequently released from the SCC, and the recidivism rate of persons released from the SCC. Dr. Leslie Rawlings, a psychologist who had interviewed Post, discussed Post’s performance in the first two phases of SCC treatment, as well as the number of additional phases of SCC treatment. Additionally, in its direct examination of Post, the State elicited testimony that he was in the second of six phases of SCC treatment and had not yet completed it. Following this line of questioning, the jury submitted a question to Post asking him where he was in the SCC treatment program.
¶6 In its opening and closing arguments, the State emphasized the benefits to Post and the community of SCC treatment, going so far as to directly compare recidivism rates of those who completed treatment in a secure facility and those who did not. The State concluded its closing remarks by stating, “Now Mr. Post’s best chance of reducing his risk before he’s released is to complete the treatment *308program at the SCC.” Verbatim Report of Videotaped Proceedings (VRVP) (Dec. 14, 2004) at 196. After the defense objected and the court instructed the jury that the jury must decide the case based on the required elements, the State reiterated that “Charles Post’s best chance to reduce his chance for recidivism is to stay in a secure facility and complete the treatment program.” Id.
¶7 After the jury found that Post was an SVP, the court ordered him civilly committed pursuant to former RCW 71.09.060(1) (2001). A divided panel of the Court of Appeals, however, reversed and remanded for a new trial. In re Det. of Post, 145 Wn. App. 728, 732-33, 187 P.3d 803 (2008). The court held that “[e]vidence concerning the final four phases of the SCC program, in which Post did not participate, and evidence regarding potential future less restrictive alternatives to total confinement were not relevant to the jury’s determination of whether Post was proved to be an SVP,” id. at 742, and further held that admission of such evidence was not harmless error, id. at 748-49. Though this resolved the case, the court proceeded to address other issues raised on appeal “in order to facilitate proceedings on remand,” id. at 749, including the exclusion of evidence about the possibility of the State’s filing a recent overt act petition.
¶8 The State filed a petition for review of the Court of Appeals decision. In his answer, pursuant to RAP 13.4(d), Post sought review of the trial court’s exclusion of evidence that he could be subject to another SVP commitment petition if he committed a recent overt act. We accepted review of the decision by the Court of Appeals, including the recent overt act petition issue. In re Det. of Post, 166 Wn.2d 1033, 217 P.3d 782 (2009).
ISSUES
¶9 1. At a trial to determine whether a person is an SVP, may the State introduce evidence of treatment that would be available to that person if committed and of potential future release to a less restrictive alternative?
*309¶10 2. Is admission of the irrelevant evidence at issue here harmless error?
¶11 3. Is evidence that a person could be subject to a future SVP commitment petition upon commission of a recent overt act relevant to the SVP determination trial?
ANALYSIS
A. Standard of Review
¶12 A trial court’s ruling on the admissibility of evidence is reviewed for abuse of discretion. City of Auburn v. Hedlund, 165 Wn.2d 645, 654, 201 P.3d 315 (2009). A trial court abuses its discretion if the “exercise of its discretion is manifestly unreasonable or based upon untenable grounds or reasons.” State v. Powell, 126 Wn.2d 244, 258, 893 P.2d 615 (1995).
B. Background
¶13 The sexually violent predator act, chapter 71.09 RCW, permits the State, in limited circumstances, to civilly commit persons determined to be SVPs. The legislature has set forth by statute detailed procedures addressing, among other things, when the State may file a petition to civilly commit a person, RCW 71.09.030; the evaluations to which the respondent may be subjected, RCW 71.09.040(4), .025(l)(b)(v); and the hearings and trials to which a respondent is entitled, RCW 71.09.040, .050, .060. These precommitment procedures culminate in an SVP determination trial. RCW 71.09.050, .060.1
¶14 At the SVP determination trial, there is but one question for the finder of fact: Has the State proved, beyond a reasonable doubt, that the respondent is an SVP? RCW 71.09.060(1). To answer this question in the affirmative, the *310jury must determine three elements: (1) that the respondent “has been convicted of or charged with a crime of sexual violence,” (2) that the respondent “suffers from a mental abnormality or personality disorder,” and (3) that such abnormality or disorder “makes the person likely to engage in predatory acts of sexual violence if not confined in a secure facility.” RCW 71.09.020(18); see also In re Det. of Audett, 158 Wn.2d 712, 727, 147 P.3d 982 (2006) (quoting In re Det. of Thorell, 149 Wn.2d 724, 758-59, 72 P.3d 708 (2003)). This third element is a compound determination that requires finding both causation (i.e., the abnormality or disorder causes the likelihood of future acts), Thorell, 149 Wn.2d at 736, 742, and that “the probability of the defendant’s reoffending exceeds 50 percent,” In re Detention of Brooks, 145 Wn.2d 275, 298, 36 P.3d 1034 (2001), overruled on other grounds by Thorell, 149 Wn.2d at 753. If the jury determines that the respondent is an SVP, the consequence, which is set forth by statute, is the civil commitment of the respondent. RCW 71.09.060(1).
C. Evidence of Postcommitment Treatment Availability Is Irrelevant
¶15 The State argues that the Court of Appeals erred by holding that evidence about treatment available at the SCC but not yet undergone by the respondent is inadmissible at the SVP determination trial. The question before this court is a narrow one. The Court of Appeals held that evidence about the content of the SCC treatment voluntarily undergone by Post at the time of trial and Post’s performance in treatment were relevant and admissible. Post, 145 Wn. App. at 741-42. Post did not appeal this holding. Our review is therefore limited to the admissibility of evidence about the content of SCC treatment programs in which Post had not yet participated and the admissibility of evidence that Post could later be released to a less restrictive alternative if he were committed. Because that disputed evidence was not relevant to the jury’s determination, we affirm the Court of Appeals.
*311¶16 It is a fundamental rule of evidence that “[e]vidence which is not relevant is not admissible.” ER 402. To be relevant, evidence need only have “any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.” ER 401.
¶17 If the disputed evidence is relevant, it must be to the second or third elements of the SVP definition. The State does not argue, nor could it, that treatment not yet undergone but available to committed SVPs makes any more probable the existence of convictions for crimes of sexual violence. Our inquiry is properly limited, then, to whether the existence and content of treatment programs at the SCC, as well as the possibility of future release to a less restrictive alternative, make it any more or less probable that Post has a mental abnormality or personality disorder that makes him “likely to engage in predatory acts of sexual violence if not confined in a secure facility.” RCW 71.09.020(18). On the face of the question, the answer plainly appears to be “no.” Post’s opportunities if committed have no obvious bearing on his actions if he is not committed, as those two potential realities are mutually exclusive.
¶18 This is not the first time we have been called upon to determine whether the consequences of commitment are relevant to the determination of whether a person is an SVP. In In re Detention of Turay, 139 Wn.2d 379, 986 P.2d 790 (1999), a respondent to an SVP determination trial sought to introduce evidence about the conditions of confinement at the SCC. Id. at 403. Affirming the trial court’s motion in limine to exclude such evidence, we made it explicit that in an SVP determination trial, “[t]he trier of fact’s role ... is to determine whether the [respondent] constitutes an SVP; it is not to evaluate the potential conditions of confinement.” Id. at 404. That statement applies with equal force in the present case. If the conditions of confinement are not relevant when offered by the respondent to an SVP proceeding, they are not relevant when offered by the State.
*312¶19 We are not alone in concluding that the consequences of a finder of fact’s determination that a person is an SVP are not relevant to making that determination. In People v. Rains, 75 Cal. App. 4th 1165, 89 Cal. Rptr. 2d 737 (1999), the California Court of Appeal addressed a similar question. In Rains, as here, the respondent to the SVP determination trial conceded the existence of prior sexually violent offenses, so the only two issues for the jury were whether Rains had “ ‘a diagnosed mental disorder’ and . . . if so, whether that disorder ‘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.’ ” Id. at 1168 (quoting Cal. Welf. & Inst. Code § 6600(a)). The trial court admitted evidence that if the jury found Rains to be an SVP, he would be committed to a hospital and receive treatment. Id. at 1167. The Court of Appeal, however, held that this evidence had no relevance to either of the issues before the jury. Id. at 1170. This is the same conclusion we reach today.
¶20 The State advances three arguments in favor of the disputed evidence’s relevance: (1) it is expressly permitted by the SVP definition’s inclusion of the phrase “if not confined in a secure facility,” (2) it is probative of Post’s mental condition and dangerousness, and (3) it is relevant to rebut Post’s voluntary treatment plan. The first of these arguments is easily rebutted. The language “if not confined in a secure facility” appears in RCW 71.09.020(18), which defines the term “ ‘[s]exually violent predator.’ ” In context, it is part of the requirement that the person’s mental abnormality or personality disorder “makes the person likely to engage in predatory acts of sexual violence if not confined in a secure facility.” Id. Far from making the conditions of a secure facility relevant, this clause operates to define the relevant inquiry as not including such conditions. The opposite reading would be painfully strained and violate our duty to strictly construe statutes involving deprivations of liberty. In re Cross, 99 Wn.2d 373, 379, 662 P.2d 828 (1983).
*313¶21 The State next argues that the treatment available at the SCC is relevant to the determination of Post’s mental state and dangerousness. Insofar as the State’s argument pertains to SCC treatment already undergone by Post, that question is not before us and we assume, as the Court of Appeals held, that such evidence is relevant and otherwise admissible. As for the phases of treatment not yet undergone by Post, the fact that those phases are available at the SCC does not make it any more or less probable that Post has a mental abnormality or a personality disorder or that he is dangerous. The same is true for potential future release from the SCC to a less restrictive alternative. With respect to dangerousness, the question for the finder of fact is whether Post is likely to engage in a predatory act of sexual violence if released immediately; it is not whether Post would be more likely to commit such an act if immediately released than if he were confined and subsequently released.
¶22 The State’s final argument in favor of the SCC treatment program evidence is that it was necessary to rebut the efficacy of Post’s voluntary treatment plan. Post’s plan for unconditional release into the community, which he presented to the jury, consisted of regular individual treatment and marital counseling with Dr. Sally Wing and group therapy with one of her colleagues. He would be living with his wife in the downstairs area of his mother’s home, working for a family friend, keeping a daily log of all his activities, and continuing to participate in a nonviolent communications group he had joined while in prison. In addition, numerous witnesses testified that they would support him in the community. This evidence was all relevant to the likelihood that Post would reoffend, as it has some tendency, if believed, to show that Post was less likely to do so. See Thorell, 149 Wn.2d at 751 (“Because this goes to whether the definition of SVP is met, the individual may bring [voluntary treatment] evidence in defense of commitment.”). As the Court of Appeals recognized in this case, the State was free “to attempt to discredit the efficacy of the *314proposed program and the true level of Post’s commitment to successful completion thereof.” Post, 145 Wn. App. at 744. This could be accomplished in any number of ways, including by highlighting elements of treatment the State’s experts believed were necessary for Post but not present in his treatment plan. Here, however, the State went substantially further by introducing evidence of the components of treatment that would be available to Post if he were committed. That evidence does not rebut Post’s treatment plan but instead presents an alternative to it, an alternative that is irrelevant to whether Post’s plan will be successful. It may be that commitment is more likely to prevent Post from committing another predatory act of sexual violence than is Post’s voluntary treatment program, but this is entirely irrelevant to the likelihood that Post will reoffend if unconditionally released.
¶23 In sum, it is manifest that the trial court abused its discretion in allowing testimony about the treatment that would be available to Post if he were committed and the possibility of release to a less restrictive alternative following commitment. The State’s arguments that the evidence is relevant are unavailing, and the admission violated ER 402 and Turay.
D. The Admission of Irrelevant Evidence Was Not Harmless Error
¶24 The improper admission of evidence about SCC treatment not undergone by Post and the possibility of future release to a less restrictive alternative was not harmless in this case. An evidentiary error is not harmless “if, ‘within reasonable probabilities, had the error not occurred, the outcome of the trial would have been materially affected.’ ” State v. Neal, 144 Wn.2d 600, 611, 30 P.3d 1255 (2001) (quoting State v. Smith, 106 Wn.2d 772, 780, 725 P.2d 951 (1986)).
¶25 The reasonable probability that the erroneous admission of evidence affected the outcome of this case is apparent. In Post’s first commitment trial, at which the *315State did not introduce evidence about SCC treatment, the jury deadlocked. In the second trial, however, the State did introduce such evidence and the jury determined that Post was an SVP. This is persuasive evidence that the introduction of the evidence may have impacted the outcome. Moreover, the evidence of treatment available at the SCC offered in the second SVP determination trial was not merely in passing, but was thorough, systematic, and repeated. Another objective example of the way in which the improper evidence may have affected the trial is that the jury submitted several questions to witnesses concerning the treatment available at the SCC. From this we may conclude that the inadmissible evidence was on the minds of jurors.
¶26 Though it is conceivable that the erroneous admission might have been cured by an appropriate limiting instruction, the trial court declined to give the limiting instruction proposed by Post that would have instructed the jury not to consider whether commitment of Post to the SCC and potential future conditional release was preferable to Post’s unconditional release and voluntary treatment plan. In its closing argument, the State directly invited that irrelevant comparison by stating, “Mr. Post’s best chance of reducing his risk before he’s released is to complete the treatment program at the SCC.” VRVP (Dec. 14, 2004) at 196. Based on all this, it is well within the range of reasonable probabilities that the improper introduction of the disputed evidence materially affected the outcome of this closely contested case.2 The error is therefore not harmless.
*316 E. Recent Overt Act Petition Evidence Is Relevant
¶27 In his answer to the State’s petition for review, Post raised the additional issue of the trial court’s exclusion of evidence that Post could be subject to a new SVP commitment petition if he committed a recent overt act. A “ ‘[r]ecent overt act’ ” is “any act, threat, or combination thereof that has either caused harm of a sexually violent nature or creates a reasonable apprehension of such harm in the mind of an objective person who knows of the history and mental condition of the person engaging in the act or behaviors.” RCW 71.09.020(12). The Court of Appeals, bound by its earlier decision in State v. Harris, 141 Wn. App. 673, 174 P.3d 1171 (2007), affirmed the trial court’s exclusion of this evidence. In Harris, which addressed an identical issue, the respondent in an SVP determination trial sought to introduce evidence that the State could file a new petition if he were released and committed a recent overt act. Id. at 679. The Harris court held that such “hypothetical evidence” was not evidence of conditions that would exist upon the respondent’s release and affirmed the trial court’s exclusion of the evidence. Id. at 680. We disagree and disapprove of both the reasoning on the issue in Harris and the Court of Appeals’ reliance upon it in this case.
¶28 Evidence that a respondent in an SVP proceeding who is subsequently released could be subject to another SVP proceeding if he commits a recent overt act is relevant and is a condition that would exist upon placement in the community. The legislature has provided that a prosecuting attorney may bring a petition to civilly commit “a person who at any time previously has been convicted of a sexually violent offense and has since been released from total confinement and has committed a recent overt act.” RCW 71.09.030(l)(e). By its plain terms, this would apply to Post if he were released and committed a recent overt act.3 Post’s knowledge of the consequences for engaging in such con*317duct may well serve as a deterrent to such conduct and, therefore, has some tendency to diminish the likelihood of his committing another predatory act of sexual violence. This likelihood, of course, is an element that the jury must address. See RCW 71.09.020(18). The possibility of a recent overt act petition is therefore relevant to the determination of whether Post is an SVP. Moreover, it is, in a literal sense, a condition to which Post would be subject if released and does not run afoul of RCW 71.09.060(1). That the filing of a new petition is not certain to occur does not make the possibility irrelevant.
¶29 We do not decide whether the evidence was admissible; we merely correct the Court of Appeals’ misapprehension and hold that the evidence is relevant and does not violate RCW 71.09.060(1). ER 403 issues of unfair prejudice and confusion of the issues are best addressed in the first instance by the trial court, subject to review for abuse of discretion.
CONCLUSION
¶30 We hold that the trial court abused its discretion in admitting irrelevant evidence about the treatment that would be available to Post if he were committed as an SVP and the possibility of future release to a less restrictive alternative. This error was not harmless. We therefore affirm the judgment of the Court of Appeals and remand the case to the trial court for further proceedings. On retrial, evidence about the possibility of a recent overt act petition should be considered relevant.
C. Johnson, Alexander, Sanders, Chambers, Fairhurst, and J.M. Johnson, JJ, concur.Cumulatively, we refer to the preeommitment aspects of chapter 71.09 RCW as an “SVP commitment proceeding.” Because the trial contemplated by RCW 71.09.050 and .060 is solely concerned with determining whether or not the person is an SVP, we employ the more descriptive term “SVP determination trial” to refer to that component.
The State also argues that Post opened the door to the testimony, but this is meritless. The trial court had already denied Post’s motions in limine to exclude such evidence and a ‘lawyer who introduces preemptive testimony only after losing a battle to exclude it cannot be said to introduce the evidence voluntarily.” State v. Vy Thang, 145 Wn.2d 630, 648, 41 P.3d 1159 (2002). Post thus did not waive his right to challenge the evidence on appeal.
We need not and do not decide the precise contours of res judicata and collateral estoppel in these circumstances, but we note that courts of other *317jurisdictions have already addressed the issue. See Turner v. Superior Court, 105 Cal. App. 4th 1046, 130 Cal. Rptr. 2d 300 (2003) (holding that a subsequent petition may be filed upon a showing of a change in circumstances); Commonwealth v. Chapman, 444 Mass. 15, 825 N.E.2d 508 (2005) (same).