Sohappy v. Board of Parole

• Court: Court of Appeals of Oregon
• Decided: November 8, 2023
• Docket: A174855
• Status: Published

Opinion

28 November 8, 2023 No. 576

IN THE COURT OF APPEALS OF THE STATE OF OREGON

DOMINIQUE JERMAINE SOHAPPY, Petitioner, v. BOARD OF PAROLE AND POST-PRISON SUPERVISION, Respondent. Board of Parole and Post-Prison Supervision A174855

Submitted November 21, 2022. Ernest G. Lannet, Chief Defender, Criminal Appellate Section, and Stephanie J. Hortsch, Deputy Public Defender, Office of Public Defense Services, filed the briefs for petitioner. Ellen F. Rosenblum, Attorney General, Benjamin Gutman, Solicitor General, and Jeff J. Payne, Assistant Attorney General, filed the brief for respondent. Before Aoyagi, Presiding Judge, and Lagesen, Chief Judge, and Jacquot, Judge. AOYAGI, P. J. Reversed and remanded. Cite as 329 Or App 28 (2023) 29 30 Sohappy v. Board of Parole

AOYAGI, P. J. Petitioner seeks judicial review of a 2020 order of the Board of Parole and Post-Prison Supervision that set petitioner’s sex offender notification level (SONL) at Level 2 (Moderate). Under ORS 163A.100, the board must “adopt by rule a sex offender risk assessment methodology for use in classifying sex offenders,” and application of that rule “must result in placing the sex offender in one of” three notifica- tion levels based on their risk of reoffending—with Level 3 offenders presenting the highest risk, Level 2 offenders presenting a moderate risk, and Level 1 offenders present- ing the lowest risk. In response to that directive, the board adopted OAR 255-085-0020, and it applied the version of that rule in effect on April 29, 2020, to petitioner. OAR 255- 085-0020 (Apr 29, 2020) provides that the board “shall use the Static-99R actuarial instrument on the Board’s website at http://www.oregon.gov/BOPPPS along with attending rules and research found on http://www.static99.org/, to con- duct a sex offender risk assessment” and place each regis- trant into one of the three notification levels. Petitioner contends that the board misinterpreted OAR 255-085-0020(1) (Apr 29, 2020) when it used the Static- 99R actuarial instrument without the attending rules and research on the Static-99R website to determine his risk level, specifically the attending rules and research related to sex-offense-free time in the community. Petitioner argues that the board’s failure to account for his 12 years of sex-of- fense-free time in the community resulted in SONL misclas- sification. The board responds that it is discretionary under the Static-99R methodology whether to consider sex-offense- free time. We agree with petitioner that the board’s inter- pretation is implausible, and that the only plausible inter- pretation of the rule required the board to use the attending rules and research on the Static-99R website regarding sex- offense-free time in the community in setting petitioner’s risk level. We therefore reverse and remand. All references to OAR 255-085-0020 in this opinion are to the version that went into effect on April 29, 2020, which is the version that the board applied to petitioner. The Cite as 329 Or App 28 (2023) 31

rule has since been amended twice, but those amendments are not at issue in this review proceeding.1 I. FACTS When petitioner was 18 years old, he engaged in sex- ual misconduct toward fellow students at the Oregon School for the Deaf, which led to his 2008 conviction for offenses requiring him to register as a sex offender. Petitioner was sentenced to 60 months’ probation, which he successfully completed, and never served any prison time. In April 2020, petitioner petitioned the board for relief from registering as a sex offender. See ORS 163A.125(1) (allowing people classified in Level 1 to request relief from the sex offender registration requirement). Because peti- tioner had never been classified under the current SONL system, the board used the Static-99R actuarial instru- ment to conduct a risk assessment and set his initial SONL under OAR 255-085-0020. See Or Laws 2013, ch 708, § 7, compiled as a note after ORS 163A.110 (addressing SONL classification of “existing registrants,” i.e., people for whom the event that triggered their obligation to make an initial report as a sex offender, such as release into the community, occurred before January 1, 2014); Baker v. Board of Parole, 305 Or App 814, 817, 473 P3d 83, rev den, 367 Or 290 (2020) (explaining that, under the current SONL system, a person convicted of a sex crime must be classified by the board to determine the intensity of the person’s reporting obligation). The board determined that petitioner’s Static-99R score was “5” and, based solely on that score, issued an order in June 2020 classifying him as Level 2 (Moderate). Petitioner requested review, asserting, as relevant here, that the board’s approach failed to take into account the 12 years that he had been living sex-offense-free in the community. As explained more later, the attending rules and research 1 The current rule requires agencies to “use the Static-99R actuarial instru- ment with the coding manual” to conduct assessments, “except as to where it conflicts with” a rule provision requiring registrants to be classified into Level 3 or Level 2 “if an assessment under OAR 255-085-0020(2) as it was at the time of release from the index sexual offense” would have resulted in that clas- sification, “without considering as part of the risk assessment the reduction of risk due to time offense-free in the community.” OAR 255-085-0020(2); OAR 255-085-0020(6). 32 Sohappy v. Board of Parole

on the Static-99R website address sex-offense-free time in the community as relevant to a sex offender’s risk of reof- fending. In early September 2020, the board rejected peti- tioner’s objections and issued a “Verification of Static-99R Score and Final Order for Sex Offender Notification Level Classification.” The board affirmed its decision to set peti- tioner’s SONL at Level 2 (Moderate), based solely on his “5” score on the Static-99R. As described in the notice sent to petitioner, that decision was “final” as to petitioner’s initial SONL classification and was “not subject to administra- tive review under OAR 255-080” but was subject to judicial review under ORS 144.335. Petitioner filed a timely petition for judicial review of the board’s SONL order. Because petitioner had petitioned for relief from reg- istration, the board also issued a second final order, denying relief from registration based on his Level 2 classification. See ORS 163A.125(1)(a) (only people classified in Level 1 are eligible to request relief from registration). On administra- tive review of that order, petitioner continued to challenge the board’s approach of setting his SONL without taking into account his 12 years of sex-offense-free time in the commu- nity. The board stated in its administrative review order, “For consideration of offense-free time in the community, the Board shall consider it when the Board conducts a reclas- sification hearing.” Petitioner filed an amended petition for judicial review to add the board’s order denying relief from registration. Petitioner seeks review of both orders, contending that the risk assessment methodology adopted by rule in OAR 255-085-0020—i.e., the Static-99R actuarial instru- ment and the attending rules and research on the Static- 99R website—required the board to consider sex-offense- free time in the community in setting his SONL. Had the board done so, petitioner asserts, he would have been classi- fied in Level 1, which would have both reduced the intensity of his reporting obligation and required the board to proceed to considering whether to relieve him from the registration requirement. A person classified in Level 1 and otherwise eligible may be relieved from the registration requirement if the board “determines, by clear and convincing evidence,” Cite as 329 Or App 28 (2023) 33

that the person “[i]s statistically unlikely to reoffend” and “[d]oes not pose a threat to the safety of the public.” ORS 163A.125(4)(a). II. PROCEDURAL ISSUES As a preliminary matter, the board argues that petitioner waived his objection, failed to exhaust his admin- istrative remedies, and failed to preserve the issue raised on judicial review. We disagree and conclude that the issue is properly before us. OAR 255-085-0040(1) provides that, with respect to SONL classification orders, “[w]ritten objections are lim- ited to presenting factual evidence regarding the Static-99R score and must be plain, concise, and directly related to specific items on the Static-99R that the registrant claims were not scored correctly.” Consistent with that rule, the board’s initial order setting petitioner’s SONL was accom- panied by a Notice of Rights that advised petitioner that he had “the right to present written factual evidence to show that [his] Static-99R score is incorrect, as explained on the Written Objections form and in OAR 255-085-0040,” and a Written Objections form that stated that objections had to be “directly related to specific items on the Static-99R.” The Written Objections form listed the 10 factual items used to calculate the Static-99R score and required petitioner to check off the items that he was claiming “were scored incor- rectly” and then provide a written explanation for each item as to why he believed it was scored incorrectly.2 The rule, the notice provided to petitioner, and the objection form provided to petitioner all indicate that the only objections to the SONL order that could be raised to the board were objections to the scoring of individual Static-99R items. Consequently, it is not at all clear that the board pro- vided a process for petitioner to challenge the board’s inter- pretation of OAR 255-085-0020(1) as reflected in its order setting his SONL at Level 2. See Golden Rule Farms v. Water Resources Dept., 321 Or App 43, 48, 515 P3d 908 (2022) (gen- erally, when “an agency provides a process for raising issues

2 The 10 factual items listed on the Written Objections form correspond to the 10 factual items used to calculate a Static-99R score. See 329 Or App at 42. 34 Sohappy v. Board of Parole

to it, the doctrine [of administrative exhaustion] requires a party to present the issue to the agency through that pro- cess before a court will consider it”); cf. Peeples v. Lampert, 345 Or 209, 219, 191 P3d 637 (2008) (“when a party has no practical ability to raise an issue,” “the preservation require- ment gives way entirely”). Nevertheless, petitioner did raise the issue to the board. On the Written Objections form, petitioner did not check any of the boxes listed, instead providing a two-page letter. Petitioner did “not claim that he was scored incor- rectly on any of the Static-99R items” but, as relevant here, included a paragraph arguing that the board should have taken into account his sex-offense-free time in the com- munity. He argued that the board’s approach used a score establishing his recidivism risk in 2008 to set his risk level in 2020. He pointed out that the Static-99R coding rules and attending research indicate that, for each five years of sex- offense-free time in the community, a past offender’s like- lihood of recidivism decreases by approximately half. He concluded by asserting that, if the board had applied the Static-99R correctly, i.e., taken into account his sex-offense- free time in the community, it would have scored him at “roughly 0, placing him at Level 1.” We disagree with the board that petitioner waived his right to challenge how the board set his risk level when he stated in his letter accompanying his written objections that “he does not claim that he was scored incorrectly on any of the Static-99R Items.” That argument takes petition- er’s statement out of context. In context, it is apparent that no waiver occurred. We also disagree with the board that petitioner failed to exhaust his administrative remedies, or failed to preserve the issue, by not raising it in the first paragraph of his letter. Petitioner’s letter is somewhat dis- organized, but he raised and developed the argument later in the letter, and the board has not identified any rule that petitioner violated by structuring his letter as he did. The fact that the board did not provide a clear mechanism to raise the issue also weighs against taking an overly strict view of how the argument was presented. Cite as 329 Or App 28 (2023) 35

As for the second order, as described above, after the board entered the SONL order, it entered a second order denying relief from registration on the basis that peti- tioner was classified in Level 2, and only people classified in Level 1 are eligible for relief from registration. Petitioner sought, and obtained, administrative review of that order. On administrative review, he renewed his argument that the board had incorrectly classified him in Level 2 because OAR 255-085-0020 required the board to take into account his sex-offense-free time in accordance with the Static-99R attending rules and research. The board rejected that argu- ment, stating that it would consider sex-offense-free time “when the Board conducts a reclassification hearing.” The board does not make any procedural arguments specific to its second order, but, in the interests of completeness, we note that petitioner exhausted administrative review as to the second order, and he now seeks judicial review of that order as well. III. STANDARD OF REVIEW We turn to the merits. Petitioner contends that the board misinterpreted OAR 255-085-0020(1) as allowing the board to rely entirely on petitioner’s Static-99R score to determine his recidivism risk, without taking into account sex-offense-free time in the community as provided in the attending rules and research on the Static-99R website. We therefore review the board’s order to determine whether the board “erroneously interpreted a provision of law.” ORS 183.482(8)(a); ORS 144.335(3). An agency’s interpretation of its own rule is entitled to judicial deference “if that interpretation is plausible given the wording of the rule, its context, and any other source of law.” OR-OSHA v. United Parcel Service, Inc., 312 Or App 424, 434, 494 P3d 959 (2021). If the agency’s interpretation is implausible, we interpret the rule using our usual con- struction methodology. County of Klamath v. Ricard, 317 Or App 608, 612, 507 P3d 333 (2022); see also Noble v. Dept. of Fish and Wildlife, 355 Or 435, 459, 326 P3d 589 (2014) (rejecting an agency’s implausible interpretation of its rule and remanding for the agency to apply the only plausible interpretation); Don’t Waste Oregon Com. v. Energy Facility 36 Sohappy v. Board of Parole

Siting, 320 Or 132, 142, 881 P2d 119 (1994) (observing that an agency’s interpretation is “erroneous” for purposes of ORS 183.482(8)(a) if it is “inconsistent with the wording of the rule itself, or with the rule’s context, or with any other source of law”). “That is, we consider the text of the rule and its context, including other portions of the rule and related laws, and the rule’s adoption history.” County of Klamath, 317 Or App at 612 (internal quotation marks omitted); see also State v. Gaines, 346 Or 160, 171-72, 206 P3d 1042 (2009) (setting out method of statutory interpretation). “[O]ur role in interpreting rules, as in interpreting statutes, is to get the answer as correct as we can.” OR-OSHA, 312 Or App at 435. IV. BACKGROUND Given the nature of the parties’ arguments, a sig- nificant amount of background information is necessary to lay the foundation for our analysis. We begin by summariz- ing the history of Oregon’s sex offender registry and, in par- ticular, the 2013 statutory changes that created the current three-tiered SONL structure. We then describe the role of the Static-99R in SONL classification. A. Oregon’s Sex Offender Registry The legislature established Oregon’s sex-offender registry in 1989. Or Laws 1989, ch 984. The registry’s purpose has always been “to assist law enforcement agen- cies in preventing future sex offenses.” ORS 163A.045(1). The registration requirement is regulatory in nature, not punitive. State v. McNab, 334 Or 469, 481-82, 51 P3d 1249 (2002) (rejecting an ex post facto challenge to the registra- tion requirement, based on the statutory purpose and the fact that “[t]he operation of the law conforms to the legisla- ture’s declared purpose”); see also, e.g., State v. Benson, 313 Or App 748, 770, 495 P3d 717, rev den, 369 Or 69 (2021) (“The requirement that an offender acknowledge that they are aware of their registration requirements exists as part of a noncriminal regulatory framework, is individually reg- ulatory in nature, and serves the noncriminal purpose of effectuating Oregon’s sex offender registration system.”). Cite as 329 Or App 28 (2023) 37

The provision that is now ORS 163A.100 was enacted in 2013. Or Laws 2013, ch 708, § 1.3 Stakeholders in Oregon’s sex offender registry—including the Oregon State Police (which administers the registry), the Oregon Department of Corrections (DOC), community corrections agencies, the Oregon Criminal Defense Lawyers’ Association, the Oregon State Sheriffs Association, and the board—were concerned that the registry had become so large that it was no longer serving its statutory purpose of assisting law enforcement agencies in preventing future sex offenses. See Testimony, House Judiciary Committee, HB 2549, Feb 28, 2013, Ex 4, at 1-2 (statement of Jeff Wood, Parole & Probation Division Commander, Marion County Sheriff’s Office) (describing origin of bill). More than 19,000 people were on the registry in 2013, and there was no way for law enforcement or super- visory agencies to identify who posed a high risk of commit- ting new sex offenses so that they could direct their limited resources toward those offenders. See id. at 1 (describing Oregon’s registry as “a bit of a Pandora’s Box” due to piece- meal legislation); Audio Recording, House Committee on Judiciary, HB 2549, Feb 28, 2013, at 29:30 (comments of Vi Beatty, Manager, Sex Offender Registry, Oregon State Police), https://olis.oregonlegislature.gov/liz/mediaplayer/?clien- tID=4879615486&eventID=2013021184 (accessed Oct 5, 2023) (likening Oregon’s constantly growing registry to a speeding battleship that urgently needed to be stopped but would take some time to stop because, even if the proposed legislation led to the removal of 5,000 registrants in the five years after passage, it would take seven to 10 years to stop the registry’s growth4). As described by Brielyn Atkins, an advocate for vic- tims of domestic violence, in a letter circulated on the house floor by Representative Jennifer Williamson, one of the bill’s sponsors, “Having a sex offender registry is important for 3 Or Laws 2013, ch 708, § 1, was codified as former ORS 181.800. As we will discuss, the provision was amended in 2015. Or Laws 2015, ch 820, § 1. The same year, former ORS 181.800 was renumbered as ORS 163A.100. 4 To the extent that the 2013 legislation was intended to stop or reverse the registry’s growth trend, it has not done so. Since 2013, the registry has grown to over 33,000 registrants. Frequently Asked Questions, Oregon State Police, Sex Offender Registry Section, Sexoffenders.Oregon.Gov/FAQ (accessed Oct 16, 2023). 38 Sohappy v. Board of Parole

victim and community safety, but having an overly broad registry can tax limited law enforcement resources by requiring unnecessary supervision of certain offenders who are not predatory and do not pose a high risk to the commu- nity.” Floor Letter, Rep Jennifer Williamson, July 6, 2013, HB 2549. The 2013 legislation was meant to address that prob- lem by creating a three-tiered registry that would “identify an appropriate level of supervision” for each offender, “which enables better allocation and use of limited resources.” Testimony, Joint Committee on Ways and Means, Public Safety Subcommittee, HB 2549, May 13, 2013, at 2 (state- ment of Jeff Wood). Everyone on the registry would be clas- sified as either a (1) “level one sex offender who presents the lowest risk of reoffending and requires a limited range of notification”; (2) “level two sex offender who presents a mod- erate risk of reoffending and requires a moderate range of notification”; or (3) “level three sex offender who presents the highest risk of reoffending and requires the widest range of notification.” Or Laws 2013, ch 708, § 1. In addition to providing for different notification lev- els, the 2013 legislation also provided mechanisms for reg- istrants to seek reclassification to a lower notification level or relief from registration altogether. Id. at § 5. However, as a policy matter, the legislature made certain exceptions. Registrants previously classified as “predatory sex offend- ers” or “sexually violent dangerous offenders” were automat- ically classified as Level 3, see id. at § 7(2), and they, along with anyone else initially classified as Level 3, may never be reclassified lower than Level 2 or relieved from the registra- tion requirement. Id. at § 5(3)(b). Any registrant convicted of a person felony or person Class A misdemeanor since their sex-offense conviction may never be reclassified at all or relieved from registration. Id. at § 5(3)(a). Other than those policy exceptions, the legislature wanted registrants to be classified based on their statistical risk of committing another sex offense. The legislation there- fore required DOC, which was charged with classifying new registrants as they entered the community, to “adopt by rule a sex offender risk assessment tool for use in classifying sex Cite as 329 Or App 28 (2023) 39

offenders based on the statistical likelihood that an indi- vidual sex offender will commit another sex crime.” Id. at §§ 1, 2. DOC represented that it would adopt the Static-99R as the risk assessment tool. Testimony, Joint Committee on Ways and Means, Public Safety Subcommittee, HB 2549, May 13, 2013, Ex 12, at 1 (statement of Cindy Booth, Oregon Department of Corrections). As for the more than 19,000 existing registrants, the legislature directed the board to classify them by December 1, 2016. Or Laws 2013, ch 708, § 7(2). The board indicated that it would use existing Static-99R scores for approximately 16,000 registrants who already had them. Joint Committee on Ways and Means, Public Safety Subcommittee, HB 2549, May 13, 2013, Ex 4, at 7 (HB 2549 Fiscal Summary, Jay Scroggin, Oregon Board of Parole and Post-Prison Supervision). The board had used the Static- 99R, or its predecessor the Static-99, since at least 2004 to help assess whether people qualified as “predatory sex offenders.” See V. L. Y. v. Board of Parole, 338 Or 44, 46 n 1, 106 P3d 145 (2005). It was therefore a well-established sta- tistical tool, and many registrants had already been scored on it. In 2015, the legislature amended the statutory pro- vision requiring DOC to adopt “a sex offender risk assess- ment tool” to classify sex offenders based on their likelihood of committing another sex crime, Or Laws 2013, ch 708, § 1, replacing it with a requirement that the board adopt “a sex offender risk assessment methodology” for that purpose, Or Laws 2015, ch 820, § 1. The wording change from “tool” to “methodology” had to do with the fact that the Static-99R is statistically validated only for adult male sex offenders, so a different tool needed to be used for female sex offenders and juvenile sex offenders. See Testimony, Joint Committee on Ways and Means, Public Safety Subcommittee, HB 2320, Apr 20, 2015, Ex 6, at 1 (statement of Heidi Steward, Oregon Department of Corrections, discussing need to use “appro- priate methods and tools with various sex offending popula- tions” and identifying female sex offenders and juvenile sex offenders as “particularly challenging” to assess because of the original statutory text). 40 Sohappy v. Board of Parole

Thus, since 2015, ORS 163A.100 has provided: “The State Board of Parole and Post-Prison Supervision shall, in consultation with community corrections agencies, adopt by rule a sex offender risk assessment methodology for use in classifying sex offenders. Application of the risk assessment methodology to a sex offender must result in placing the sex offender in one of the following levels: “(1) A level one sex offender who presents the lowest risk of reoffending and requires a limited range of notification. “(2) A level two sex offender who presents a mod- erate risk of reoffending and requires a moderate range of notification. “(3) A level three sex offender who presents the highest risk of reoffending and requires the widest range of notification.” The board thereafter promulgated a rule as directed. The version of the rule in effect on April 29, 2020, which is the version relevant to this case, states: “For classification and community notification for adult male registrants, the classifying agency shall use the Static-99R actuarial instrument on the Board’s website at http://www.oregon.gov/BOPPPS along with attending rules and research found on http://www.static99.org/, to conduct a sex offender risk assessment. Classifying agencies may score registrants using information from previous Static-99 or Static-99R assessments. Classifying agencies shall score and place each registrant into one of the following levels: (a) Notification Level 1: Low risk; (b) Notification Level 2: Moderate risk; or (c) Notification Level 3: High risk.” OAR 255-085-0020(1). The board was unable to complete the classification of existing registrants within the timeline set by the legisla- ture, and the legislature has repeatedly extended the dead- line. Or Laws 2015, ch 820, § 27 (extending the deadline to December 1, 2018); Or Laws 2017, ch 488, § 1 (extending the deadline to December 1, 2022); Or Laws 2019, ch 430, § 1 (extending the deadline to December 1, 2026). In 2019, the board advised the legislature that, without funding for addi- tional staff, it would take over 40 years to finish classifying Cite as 329 Or App 28 (2023) 41

everyone on the registry. House Committee on Judiciary, HB 2045, Feb 4, 2019, Ex 4, at 7 (Dylan Arthur, Oregon Board of Parole and Post-Prison Supervision, HB 2045 Agency Presentation). One consequence of the delay in clas- sifying people who were already on the registry when the three-tiered system went into effect in 2014 is that some registrants—like petitioner—do not receive an initial SONL until they petition for “reclassification” or relief from registration. B. The Static-99R Because the board selected the Static-99R actu- arial instrument and attending rules and research on the Static-99R website as the risk assessment methodology to be used in classifying adult male sex offenders like peti- tioner, we next describe that methodology. The coding rules for the Static-99R—see Amy Phenix, Yolanda Fernandez, Andrew J. R. Harris, Maaike Helmus, R. Karl Hanson, & David Thornton, Static-99R Coding Rules Revised, 2016, available at https://saarna.org/static-99/ (accessed Oct 6, 2023) (Coding Rules)5 —as well as the research cited in this section were available on the Static-99R website in April 2020 and therefore were part of the “attending rules and research found on http://www.static99.org/” referenced in OAR 255-085-0020. See Capture of Static99.Org from April 15, 2020, Web.Archive.org, available at http://web.archive. org/web/20200415133800/http:/www.static99.org/ (accessed Oct 6, 2023) (showing Static-99R website as of April 15, 2020).6 The “Static-99R is an actuarial risk assessment instrument designed to assess risk of sexual recidivism for adult males who have already been charged with or convicted of at least one sex offence against a child or a non-consent- ing adult.” Coding Rules at 12. The most recent sex offense for which a person has been arrested, charged, or con- victed is the “index offense.” Id. at 38. A person is scored on

5 The Coding Rules are also available as an exhibit to the board’s current rules. OAR 255-085-0020 (Aug 16, 2022) (Exhibit STATIC-99R). We agree with the parties that the Coding Rules are “attending rules” under OAR 255-085-0020(1). 6 In late April or early May 2020, the Static-99R website moved. It is now located at https://saarna.org/static-99/. 42 Sohappy v. Board of Parole

10 factual items pertaining to their personal and criminal history at the time of the index offense or, for certain items, at the time of release for the index offense:7 (1) the person’s age at release for the index sexual offense; (2) whether the person ever lived with an intimate partner for two contin- uous years; (3) index convictions for nonsexual violence; (4) prior convictions for nonsexual violence; (5) prior sexual offenses; (6) number of prior sentencing dates; (7) convic- tions for noncontact sex offenses; (8) having any unrelated victim; (9) having any stranger victim; and (10) having any male victim. Id. at 45-84. For example, the second item “is scored based on relationship history prior to release from the index offence.” Id. at 49. The scores on each item are added together to determine the person’s Static-99R score. Id. at 85. The Static-99R predicts a person’s risk of sexual recidivism at a specific point in time, which is “on the day of their first opportunity to reoffend after the index offence (e.g., release from prison for the index sex offence, conviction date if they received a non-custodial sentence, or date of charge if there was no conviction).” Id. at 66. “No matter how much time has passed since then, the score still summarizes what their risk was like on that day.” Id. In other words, because a person’s Static-99R score is based on historical facts as they existed on the date of release for the index offense, the score will never change, regardless of when it is calculated or how much time passes. As long as the index offense remains the same, a person’s Static-99R score will be the same whether it is calculated on their release date or, for example, 12, 20, or 50 years later. According to the Coding Rules and research on the Static-99R website, a risk assessment based solely on the Static-99R score is statistically valid for “approximately two years” from release. Id. at 13. Essentially, a person’s Static- 99R score remains the same over time (absent a new index 7 In the context of the Static-99R, the date of “release” means the date when “the offender is ‘free’ (in the community) after the index sex offence is processed and therefore has an opportunity to reoffend. It may refer to release from court, jail, prison, psychiatric hospital, or the like. Offenders are considered in the com- munity if they are on parole, probation, or other types of community supervision. If they do not receive a custodial sentence for their index offence, the release date would be the date of conviction.” Coding Rules at 48. Cite as 329 Or App 28 (2023) 43

offense), while the risk of sexually reoffending changes over time, predictably declining for those who remain sex-offense- free in the community. David Thornton, R. Karl Hanson, Sharon M. Kelley, & James C. Mundt, Estimating Lifetime and Residual Risk for Individuals Who Remain Sexual Offense Free in the Community: Practical Applications, 33(I) Sexual Abuse 3 (2021), available at https://saarna.org/ research/ (accessed Oct 6, 2023) (Estimating Lifetime and Residual Risk).8 Recognizing that evaluators may want to assess a person’s sexual recidivism risk at a point in time later than two years after release, the Coding Rules address that issue. As to sex-offense-free time in the community, the Coding Rules state: “In some cases, evaluations may be for offenders who have had a substantial period at liberty in the community (since their release from the index sex offense * * *) with oppor- tunity to sexually reoffend, but have not done so. The lon- ger an offender has been free of detected sexual offending since [their] release to the community from their index sex offence, the lower their risk of recidivism. Our research has found that, in general, for every five years the offender is in the community without a new sex offence, their risk for recidivism roughly halves. Consequently, we recommend that for offenders with two years or more sex offence free in the community since release from the index offence, the time they have been sex offence free in the community should be considered in the overall evaluation of risk. Static risk assessments estimate the likelihood of recidivism at the time of release and we expect they would be valid for

8 Estimating Lifetime and Residual Risk was available on the Static-99R website before April 2020 but was formally published in 2021, so we use the 2021 date. As summarized in the article’s abstract: “Although individuals with a history of sexual crime are often viewed as a lifelong risk, recent research has drawn attention to consistent declines in recidivism risk for those who remain offense free in the community. Because these declines are predictable, this article demonstrates how evaluators can use the amount of time individuals have remained offense free to (a) extrap- olate to lifetime recidivism rates from rates observed for shorter time peri- ods, (b) estimate the risk of sexual recidivism for individuals whose current offense is nonsexual but who have a history of sexual offending, and (c) calcu- late yearly reductions in risk for individuals who remain offense free in the community.” Estimating Lifetime and Residual Risk, 33(I) Sexual Abuse at 3. 44 Sohappy v. Board of Parole

approximately two years. For offenders released for longer than two years and who have remained sex offence free, consider their overall behavior and factors external to Static-99R in your overall risk assessment.” Id. at 13-14 (internal citation omitted). The Static-99R web- site also provides a “Lifetime and Residual Risk Calculator,” with accompanying user manual, as an evidence-based method to account for sex-offense-free time in the commu- nity. David Thornton et al, Time Free in the Community Calculator, available at https://saarna.org/static-99/ (accessed Oct 6, 2023); David Thornton et al, User Manual Lifetime Residual Risk Calculator, available at https:// saarna.org/static-99/ (accessed Oct 6, 2023). The calculator is based on the research discussed in Estimating Lifetime and Residual Risk. Thornton et al, User Manual Lifetime Residual Risk Calculator at 1. It should be noted that sex-offense-free time occu- pies a unique place in the Static-99R methodology. It is the only factor external to the Static-99R score that the Coding Rules specifically recommend how to address. As to other external factors, the Coding Rules simply state that a “pru- dent evaluator will always consider other external factors * * * that may influence risk in either direction,” including “dynamic or changeable risk factors,” as well as factors such as an offender’s stated intentions to cause further harm (higher risk) or restricted ability to reoffend due to health or a structured living environment (lower risk).9 Coding Rules at 7. 9 Another “external factor” that may increase the risk of sexual recidivism is a person’s commission of non-sexual offenses since release into the community on the index sex offense. Coding Rules at 14. “[A] new conviction for post-index non-sexual offending increases risk,” and that “effect is additive to and indepen- dent from the [sex offense] time free effect.” L. Maaike Helmus et al, Static-99R & Static 2002R Evaluators’ Workbook (Sept 28, 2021) at 5, available at https:// saarna.org/static-99/ (accessed Oct 6, 2023) (Evaluators’ Workbook). If an eval- uator wishes to consider that external factor as part of a person’s risk assess- ment, there is an option to include the information in the calculator available on the Static-99R website. Thornton et al, User Manual Lifetime Residual Risk Calculator at 2 (identifying “three factors determining risk for future sexual offending” that the calculator can account for, including the “[i]ncrease in risk for individuals who reoffend non-sexually after release from the sentence served for their index sex offense”). We note that the Evaluators’ Workbook (referred to in the Coding Rules as the “Evaluators’ Handbook”) cited in the preceding paragraph is a companion Cite as 329 Or App 28 (2023) 45

Finally, the Coding Rules emphasize that strict adherence to the coding rules is critical to a statistically valid result. “The instrument’s ability to rank offenders in terms of their relative risk for sexual recidivism has been shown to be robust across many settings using a variety of samples.” Id. at 6 (internal citation omitted); see also id. at 7 (when correctly used, the Static-99R’s predicative accuracy on a scale of 0 to 1 is .69 or .70, which is “moderate predictive accuracy”). However, deviating from the coding rules under- mines the reliability of the Static-99R: “It is important to score all items according to the scor- ing rules in this coding manual. Although the coding rules may not address all possible situations (requiring some professional judgement) and there may be some situations where the coding rules seem counter-intuitive because of the nuances of a particular case, it is important to stick to these coding rules as much as possible and not to override them with your own judgement (even when strict adher- ence to the coding rules feels silly). The reason that it is necessary to stick to the coding rules as closely as possible is because the further you deviate from the rules, the less applicable the research base behind the scale will be, and the normative data from the scale (e.g., percentiles, risk ratios, and recidivism estimates) may no longer be applica- ble. In order to benefit from the evidence base that supports the use of the scale, you must use the scale in a way that is consistent with the manual.” Id. at 7. Similarly, adjusting a Static-99R score to account for external factors, rather than identifying them as exter- nal factors, “or adding ‘over-rides,’ distances Static-99R esti- mates from their empirical base and substantially reduces their predictive accuracy.” Id. V. ANALYSIS Having laid the foundation for our discussion, we turn to the specifics of this case. As previously mentioned, the board determined that petitioner’s Static-99R score was

document to the Coding Rules that provides “information on how to interpret and report the [Static-99R] score results (including both relative and absolute risk information).” Coding Rules at 4. It “is updated periodically to incorporate advances in research” and is separate from the Coding Rules due to the expec- tation that “updated research will require frequent updates to the [Evaluators’ Workbook], but not the [Coding Rules].” 46 Sohappy v. Board of Parole

“5” and, based solely on that score, classified him as Level 2 (Moderate), i.e., classified him as presenting a moderate risk of committing a new sex offense. See ORS 163A.100 (defin- ing Level 2). Petitioner argues that, had the board properly applied the Static-99R methodology, including the attending rules and research, it would have assessed him as present- ing a very low risk of committing a new sex offense and clas- sified him in Level 1 (Low). The crux of petitioner’s argu- ment is that the board has implausibly interpreted OAR 255-085-0020(1) as allowing it to disregard sex-offense-free time in the community when setting a registrant’s SONL. In petitioner’s view, the only plausible interpretation of the rule is that, when the board assesses the risk that a regis- trant will commit a new sex offense and sets their SONL to reflect that risk, the board must use the attending rules and research on the Static-99R website regarding sex-offense- free time in conducting its risk assessment. Had the board done so in this case, petitioner contends, the board would have recognized that, although petitioner presented a mod- erate risk of reoffending 12 years ago when he was released into the community (as “release” is used in the Static-99R materials), he currently presents a very low risk of reoffend- ing, according to the Static-99R rules and research. The board responds that it is not required to account for sex-offense-free time in the community when doing the risk assessment to set a registrant’s initial SONL. In the board’s view, the Static-99R methodology gives the board “discretion” whether—and how—to consider sex-offense- free time in the risk assessment. The board points to its use of an “Age Chart” to score Item 1 on the Static-99R as how it has chosen to exercise that discretion. Before we address the issues on which the parties disagree, we observe that there is a foundational issue that does not appear to be in dispute, which is that the board’s charge under ORS 163A.100 is to classify sex offenders into one of three notification levels based on their risk of reof- fending at the time of the assessment. Thus, in this case, for example, the board is not assessing the risk that petitioner presented when he was released 12 years ago or deciding Cite as 329 Or App 28 (2023) 47

what notification level would have been warranted 12 years ago—it is assessing the risk that he presents now and what notification level is warranted now.10 We agree with that tacit premise of the parties’ argument. It is supported by the text of ORS 163A.100, which requires application of the adopted risk assessment methodology to place each sex offender into the notification level that corresponds to the degree of risk that that sex offender “presents.” The use of the present tense verb “pres- ents” strongly suggests that the board is to assess current risk. That is also consistent with the purpose of the stat- ute, which is to assist law enforcement in preventing future sex offenses by classifying registrants based on their risk of committing a new sex offense. Law enforcement is concerned with present risk, not historical risk. Finally, the Supreme Court has held that the registry is regulatory in nature, not punitive, in part precisely because “[t]he operation of the law conforms to the legislature’s declared purpose.” McNab, 334 Or at 480. For all of those reasons, we agree that the board’s task in conducting risk assessments under ORS 163A.100 and OAR 255-085-0020(1) is to assess the risk that the reg- istrant presents at the time of the risk assessment. We now turn to the matters on which the parties disagree. The principal point of disagreement is whether the risk assessment methodology that the board adopted in OAR 255-085-0020(1) for use in classifying adult male sex offenders—that is, the Static-99R actuarial instrument on the Board’s website along with attending rules and research found on the Static-99R website—requires the board to 10 We recognize that, when it created the three-tiered SONL system in 2013, the legislature likely was unaware of the significance of sex-offense-free time to sexual recidivism risk and likely was not thinking about that issue with respect to existing registrants. Past sex offenders have historically often been “viewed as a lifelong risk,” Estimating Lifetime and Residual Risk, 33(I) Sexual Abuse at 3, and the research regarding sex-offense-free time was significantly less developed in 2013 than it is now. However, the fact remains that the 2013 legislature enacted a statute that tasked the board with assessing present risk. Moreover, the board promulgated a rule requiring it to use the rules and research on the Static-99R website, which are more current than what was available in 2013. Cf. Testimony, Joint Committee on Ways and Means, Public Safety Subcommittee, HB 2320, Apr 20, 2015, Ex 6, at 1 (statement of Heidi Steward, Oregon Department of Corrections, recognizing—albeit in a different context—the benefit of having a statute that is flexible enough “to allow for adoption of new or proven methods and tools as best practices evolve over time”). 48 Sohappy v. Board of Parole

account for sex-offense-free time in the community (as peti- tioner argues) or simply permits the board to consider sex- offense-free time in the community if it chooses to do so (as the board argues). We have held that, under a version of OAR 255- 085-0020 substantively the same as the one at issue here, the board has discretion in applying the Static-99R to the extent that its choices are consistent with the Coding Rules. Baker, 305 Or App at 822 (concluding that the board did not abuse its discretion in declining to contact a collateral source to confirm a fact that the petitioner reported on his Static-99R questionnaire, because the Coding Rules gave the evaluator discretion to decide whether to do so); see also Stewart v. Board of Parole, 312 Or App 32, 36, 492 P3d 1283 (2021) (“[W]e review for legal error the board’s interpreta- tion of the Static-99R Coding Rules.”).11 Relying on that principle, the board argues that “the recommendations in the Static-99R Coding Rules do not impose an affirmative obligation on the board to score an individual’s Static-99R or assess an individual differently based on offense-free time. Rather, they provide the board discretion to consider that circumstance in conducting an assessment.” In particular, the board points out that the Coding Rules “recommend” taking sex-offense-free time into account “for offenders who have had a substantial period at liberty in the community * * * with opportunity to sexually reoffend, but have not done so.” Coding Rules at 13. In isolation, the use of the word “recommend” in the Coding Rules could be read to suggest that it is the evalua- tor’s choice whether to consider sex-offense-free time in the community as part of the risk assessment. However, view- ing that statement in context, we are unpersuaded by the board’s argument. The Coding Rules make clear that the Static-99R score itself measures the risk of recidivism at a specific point in time—“on the day of [the person’s] first opportunity to reoffend after the index offence (e.g., release from prison 11 In Baker and Stewart, there were no questions about whether the board’s choices were consistent with the research found on the Static-99 website, so we did not consider or discuss that issue. Cite as 329 Or App 28 (2023) 49

for the index sex offence, conviction date if they received a non-custodial sentence, or date of charge if there was no con- viction).” Id. at 66. “No matter how much time has passed since then, the score still summarizes what their risk was like on that day.” Id. Because a person’s sexual recidivism risk declines over time as time passes without a new sexual offense, the Static-99R score is “valid,” that is, it is a statis- tically valid measure of a person’s risk of sexual recidivism, only at release and for “approximately two years” there- after. Id. at 13. Thus, if, more than two years after a per- son’s release, an evaluator uses the Static-99R score alone to assess the person’s recidivism risk, the evaluator will nec- essarily reach a statistically invalid conclusion, according to the information in the Coding Rules. Even if the Coding Rules on their own could be read to only suggest, not require, that evaluators account for sex-offense-free time in conducting risk assessments, the research found on the Static-99R website—which OAR 255-085-0020(1) expressly requires the board to “use” in conducting a sex offender risk assessment—is unequivocal that sex-offense-free time must be considered to achieve a statistically valid result. That research shows that a past sex offender’s sexual recidivism risk predictably declines over time as time passes without the commission of a new sex offense and that a Static-99R score accurately predicts sexual recidivism risk for only a limited time after release. Estimating Lifetime and Residual Risk, 33(I) Sexual Abuse at 3. Moreover, the more sex-offense-free time that an eval- uator fails to account for, the more statistically inaccurate the risk assessment will be. See id. (the risk of committing a new sex offense declines with each year that a person is free in the community and does not reoffend). In short, the Coding Rules’ explanation of the statis- tical principles underlying its rules demonstrates that, as to registrants who have been living in the community for more than two years since release, the board must consider sex- offense-free time to arrive at a statistically valid assessment of the risk that the registrant will commit a new sex offense, and the research on the Static-99 website confirms and elab- orates on that fact, including providing a statistically valid 50 Sohappy v. Board of Parole

means of calculating the effect of sex-offense-free time on sexual recidivism risk. Thus, to “use the Static-99R actuar- ial instrument on the Board’s website * * * along with attend- ing rules and research found on http://www.static99.org/, to conduct a sex offender risk assessment,” the board had to consider petitioner’s sex-offense-free time in the community. Any other interpretation of OAR 255-085-0020(1) would result in the board disregarding the attending rules and research on the Static-99R website, rather than using them. Our conclusion also is consistent with the statutory scheme that the board adopted the rule to implement. To effectuate the three-tiered registry created by the legisla- ture to help law enforcement prevent future sex crimes, and subject to certain exceptions, ORS 163A.100 tasks the board with assessing each existing and new registrant to deter- mine the statistical risk of committing another sex offense that they “present.” We have previously invalidated a rule “that does not, in fact, relate to what the board is supposed to measure.” V. L. Y., 338 Or at 53 (invalidating a DOC rule regarding the board’s classification of “predatory sex offend- ers,” where the relevant statute required DOC to “develop a scale that identifies those characteristics or combination of characteristics that ‘show a tendency to victimize or injure others’ ” and did “not authorize [DOC] to devise a scale that narrows or alters the board’s inquiry or require the board to limit its inquiry to a scale that does not, in fact, relate to what the board is supposed to measure”). In sum, we agree with petitioner that the board implausibly interpreted OAR 255-085-0020 as allowing it to disregard a registrant’s sex-offense-free time in the com- munity when using “the Static-99R actuarial instrument * * * along with attending rules and research found on http:// www.static99.org/” to assess the risk that a registrant will commit a new sex offense. The board’s approach improperly resulted in the board setting petitioner’s SONL based on his risk of reoffending during 2008 to 2010 (the two-year period after his release for the index offense), rather than his risk of reoffending in 2020, when the board assessed him and decided what risk he presents to the community at that time and the concomitant appropriate reporting level. Cite as 329 Or App 28 (2023) 51

Having concluded that the board’s rule required it to account for sex-offense-free time in the community in a manner consistent with the Static-99R actuarial instrument and the attending rules and research found on the Static- 99R website, we lastly address the board’s “Age Chart.” The board contends that using the Age Chart is a permissible way to account for sex-offense-free time in the community. We disagree. The “Age Chart” is not something available on the Static-99R website. Rather, it is a tool developed by an unknown person and used by the board in scoring Items 1 and 2 of the Static-99R. Only Item 1 is at issue in this case, so we limit our discussion to Item 1. The Coding Rules pro- vide that Item 1 is to be scored based on the person’s age at the time of release for the index sex offense, with different point amounts for different age brackets. Coding Rules at 46. A person is to receive one point if aged 18 to 34.9 years old at release, zero points if aged 35 to 39.9 years old at release, negative-one point if aged 40 to 59.9 years old at release, and negative-three points if aged 60 years or older at release. Id. The board follows the Coding Rules for Item 1 for some, but not all, registrants. If a registrant was released more than 10 years before the assessment and has not been convicted in the past 10 years of a person felony or person Class A misdemeanor, then, instead of using the registrant’s age at the time of release for the index sex offense (per the Coding Rules), the board uses the registrant’s age at the time of the assessment (per its Age Chart). See Oregon Board of Parole and Post-Prison Supervision, SONL - Age Chart, available at https://www.oregon.gov/boppps/Documents/ R%26R/S99R_AgeChart2019.pdf (accessed Oct 9, 2023) (Age Chart). Sometimes, as in petitioner’s case, the board’s use of the Age Chart has no effect on the scoring of Item 1.12 However, if a person has changed age brackets between their release date and their assessment date, the use of the 12 Petitioner was 20 years old in 2008 (his release date for Static-99R pur- poses) and was 32 years old in 2020 (his assessment date). Under the Coding Rules, he would receive one point for Item 1 because he was in the 18-to-34.9 age bracket in 2008. Relying on its Age Chart, the board scored petitioner one point for Item 1 because he was in the 18-to-34.9 age bracket in 2020. 52 Sohappy v. Board of Parole

Age Chart will reduce the person’s score on Item 1 and, thus, their overall Static-99R score. We agree with petitioner that use of the Age Chart violates the Coding Rules, particularly the rule that exter- nal factors be addressed separately from the Static-99R score, rather than being “ ‘added’ to the Static-99R score or used in any way to adjust the Static-99R score.” Coding Rules at 7; see also id. (“It is important to score all items according to the scoring rules in this coding manual. * * * In order to benefit from the evidence base that supports the use of the scale, you must use the scale in a way that is consistent with the manual.”). Use of the Age Chart is there- fore not permitted by OAR 255-085-0020(1), which requires the board to use the Static-99R actuarial instrument and attending rules and research on the Static-99R website to conduct risk assessments. OAR 255-085-0020(1) does not allow the board to use the Age Chart in place of part of the Static-99R methodology. VI. CONCLUSION To achieve its objectives in creating a three-tiered registry that is more usable by law enforcement to prevent future sex crimes, the legislature tasked the board with adopting a methodology for use in classifying sex offenders based on their risk of committing a new sex offense. The board adopted the Static-99R actuarial instrument and attending rules and research on the Static-99R website as its methodology for assessing adult male sex offenders. We agree with petitioner that the only plausible interpre- tation of the version of OAR 255-085-0020(1) in effect on April 29, 2020, is that it requires the board to conduct its risk assessments in accordance with the Static-99R rules and research, which includes rules and research regarding sex-offense-free time in the community. The board failed to comply with its own rule when it assessed petitioner. Reversed and remanded.