R v Peta
[2007] NZCA 28
•28 February 2007
For a Court ready (fee required) version please follow this link
NOTE: PUBLICATION OF NAME OR IDENTIFYING PARTICULARS OF COMPLAINANTS PROHIBITED BY S 139 CRIMINAL JUSTICE ACT 1985.
˝
IN THE COURT OF APPEAL OF NEW ZEALAND
CA48/06 [2007] NZCA 28
THE QUEEN
v
JACOB RANGAITI PETA
Hearing: 18 July 2006 and 30 November 2006
Court: Glazebrook, Chisholm and Wild JJ Counsel: G P Mason for Appellant
S B Edwards for the Crown
Judgment: 28 February 2007 at 2pm
JUDGMENT OF THE COURT
The appeal is allowed and the extended supervision order imposed on Mr Peta
is quashed.
REASONS OF THE COURT
(Given by Glazebrook J)
R V PETA CA CA48/06 28 February 2007
Table of Contents
Para No
Introduction [1] Overview of the ESO regime [5] Assessment methodology used by health assessors [15] Static-AS or Automated Sexual Recidivism Scale (ASRS) [17]
Sex Offender Needs Assessment Rating (SONAR) [31] The revised psychopathy checklist (PCL-R) and sexual deviancy [39] Other relevant factors [44]
Best practice in risk assessment [50] Why we had to deal with the matter afresh [55] The effect of the new evidence on the original health
assessor’s report [62]
Mr Peta’s ASRS score [62] The effect of the lack of previous convictions and boy victims [77] The alleged double counting of violent offences [82]
The assessment of Mr Peta performed by Dr Vess [85]
Is an ESO justified based on the evidence now before
the Court? [95]
Result [105]
Introduction
[1] Mr Peta is not what is colloquially known as a decent law abiding citizen. Even as a juvenile he had extensive criminal convictions, beginning in 1999 with an aggravated robbery at the age of 14. As an adult, Mr Peta has been convicted of a variety of offences, including theft, assault, possession of a pistol, unlawful interference with a motor vehicle, escape from custody, possession of cannabis and, the most relevant convictions for this appeal, two counts of committing an indecent act with a girl under 12.
[2] The question for this appeal is whether Mr Peta should be subject to an extended supervision order (ESO) for the next ten years due to the risk he poses of sexually re-offending against children. Judge Dawson imposed the ESO on
5 January 2006 after considering a report by a senior psychologist from the Department of Corrections. We began hearing Mr Peta’s appeal from that decision on 18 July 2006. By minute of 31 July 2006, we informed the parties that we considered that the reasons given by the Judge for the imposition of an ESO were so inadequate that the matter had to be reconsidered afresh. As an appeal against the
imposition of an ESO is treated as an appeal against sentence (see s 107R(2) of the Parole Act 2002 (the Act)) and as we have no power to remit a sentence for reconsideration, this Court was obliged to deal with the matter itself: see s 385(3) of the Crimes Act 1961.
[3] On the evidence that was before the Court in July we were not able to decide whether or not an ESO should have been made. Given this and the protective nature of the legislation (see s 107I(1) of the Act and W v Chief Executive of the Department of Corrections CA172/05 9 November 2005 at [43]), we gave the parties leave to adduce further evidence. We indicated a number of areas where we would find such evidence helpful. We heard this further evidence and updated submissions from the parties on 30 November 2006.
[4] We propose to cover the following topics in this judgment:
(a) An overview of the ESO regime;
(b) The assessment methodology used by health assessors; (c) Best practice for risk assessment;
(d) Why we had to deal with the matter afresh;
(e) The effect of the new evidence on the original health assessor’s report;
(f) Mr Peta’s ASRS score;
(g) The assessment of Mr Peta performed by Dr Vess;
(h)Whether an ESO is justified on the basis of the evidence now before the Court.
Overview of the ESO regime
[5] The ESO regime came into force on 7 July 2004. It is designed to protect the public from those who pose a real and ongoing risk of committing sexual offences
against children or young persons: see s 107I(1) of the Parole Act 2002. In order to be eligible for the imposition of an ESO, an offender must have committed and been sentenced to imprisonment for a “relevant offence” as specified in s 107B, which effectively includes all sexual offences committed in respect of persons under the age of 16 years. Application for an ESO must be made while the offender is still in prison (whether or not the latest sentence expiry date was for the relevant offence) or is subject to release or detention conditions: see s 107F(1).
[6] Before imposing an ESO, the Court has to be satisfied, after considering a health assessor’s report, that the offender is likely to commit any of the relevant offences in the future: see s 107I(2) of the Act. Section 107F(2) sets out the factors that a health assessor’s report must address, preferably directly but it may do so by inference: see Grieve v Chief Executive of the Department of Corrections (2005)
22 CRNZ 20 at [14] and [25]. The s 107F(2) factors are:
(a)The nature of any likely future sexual offending by the offender, including the age and sex of likely victims;
(b) The offender’s ability to control his or her sexual impulses;
(c) The offender’s predilection and proclivity for sexual offending;
(d)The offender’s acceptance of responsibility and remorse for past offending;
(e) Any other relevant factors.
[7] The making of an ESO is a judicial decision and not that of the health assessor: see Grieve at [15]. In Barr v The Chief Executive of the Department of Corrections CA60/06 20 November 2006 at [32], this Court warned against merely rubber-stamping health assessors’ reports. It said:
[32] We wish to make it clear, however, that first instance Judges need not accept it as necessary, or right, to rubber stamp opinions of health assessors advanced by the Department of Corrections in ESO applications (and we make it clear that the Judge did not do that in this case). What is required is a careful assessment of all the historical and current factors, along with expert opinions of others, bearing in mind that an ESO can have substantial ongoing impact on an offender who has already completed the
sentence imposed by the Court for the offending. The risk of re-offending has to be such that cannot properly be ignored when viewed against the gravity of likely re-offending. Naturally, every case requires individual judgement and assessment. Jurisdiction to make the ESO arises only if there is a real and ongoing risk of committing relevant sexual offences.
[8] The Judge’s decision to impose an ESO must be made on the basis of the test set out in s 107I(2) in light of the factors set out in s 107F(2)(a) – (d) and any other relevant factors. The effect of that statutory test is that the jurisdiction for making an ESO depends on the risk of relevant offending being both real, ongoing and one that cannot sensibly be ignored having regard to the nature and gravity of the likely re-offending: see R v Belcher CA 184/05 19 September 2006 at [11].
[9] The term of any ESO must be the minimum period required for the safety of the community in light of the level of risk posed by the offender, the seriousness of the harm that might be caused to victims and the likely duration of the risk: see s 107I(5). The term cannot exceed ten years: see s 107I(4).
[10] This Court in R v Brown (2005) 22 CRNZ 233 at [51](b) and [53] accepted the thrust of a Crown submission that, in setting the term of any ESO, the Court is not concerned with ensuring that the term is proportionate to the offences that have already been committed, but with managing the risk of future offending. While the statement is correct as far as it goes, proportionality in relation to likely future offending is a relevant factor in setting the term of an ESO in the same way that there is an element of proportionality in deciding whether or not an ESO should be imposed: see the test enunciated above at [8]. This follows from the reference in s 107I(5)(b) to the seriousness of the harm that might be caused to victims. The more serious the likely future offences, the more serious the harm would usually be to victims. We also note that the likely seriousness of future offences is usually best predicted by reference to past behaviour (as discussed below at [45]).
[11] That said, the main focus in setting the term of any ESO must be on the safety of the community and, in particular, that of children and young persons, in light of the factors set out in s 107I(5). In Belcher, at [108], this Court approved the comments of the Full Court of the High Court in Chief Executive of the Department of Corrections v McIntosh HC CHCH CRI 2004-409-162 8 December 2004 at [27]
that orders are not to be made for the minimum period required to facilitate treatment but rather for the minimum period required to achieve protection of vulnerable members of the community.
[12] The Parole Board is responsible for imposing appropriate restrictions on those subject to ESOs. The most draconian possible condition appears in s 107J(1)(b). This provides that the Parole Board may impose a special condition that a person must live at a specified address and be subject to the standard detention conditions set out in s 36(2)(a), (b) and (d) of the Act as if the person were on home detention. Under s 107K(2)(a), a requirement can be imposed that the offender be accompanied and monitored for up to 24 hours a day by a person approved by the Chief Executive of the Department of Corrections. The equivalent of home detention can only apply within the first 12 months: see s 107K(3)(b). Under s 107J(1), if an offender is not subject to the equivalent of home detention, the standard release conditions apply (see s 14 of the Act), together with any special conditions (see s 15 of the Act) imposed by the Parole Board.
[13] As the imposition of ESOs through the criminal justice system involves significant restrictions (including detention) on offenders and they are imposed in response to criminal behaviour, ESOs amount to punishment: see Belcher at [49]. Given this, it is perhaps surprising that more offenders have in the past not called their own evidence with regard to s 107F(2) factors (particularly of the individualised risk factors) when they seek to challenge the imposition of an ESO.
[14] An offender can apply for variation or discharge of any condition of an ESO under s 107O(1). The Crown submitted that the power to apply for discharge or variation of conditions under s 107O(1) provides a safeguard for the offender whose risk has diminished or whose risk may have been wrongly assessed. While this is correct, it does not justify any watering down of the criteria for the imposition of an ESO. Because the imposition of an ESO has the potential to impose major restrictions on a person’s freedom of movement and association, the Court must be satisfied that the test set out above at [8] is met at the time of imposition.
Assessment tools used by health assessors
[15] This section is based on the evidence of Mr Riley, called by the Crown, who is the director of the Psychological Service of the Department of Corrections, and Dr Vess, called by Mr Peta, who is currently a senior lecturer in abnormal psychology and adult mental health at Victoria University. Both men have extensive experience in assessing and treating sexual offenders, in Dr Vess’ case both in California and New Zealand.
[16] Mr Riley and Dr Vess were in general agreement that empirically validated actuarial measures for ascertaining risk based on static factors should form the foundation of risk assessments in relation to sex offenders. Various changeable or dynamic factors, any other risk factors such as psychopathy or sexual deviancy and other aetiological (causative) factors should then be considered in formulating an individualised assessment of the nature of the risk presented by an individual offender. Risk assessments using actuarial tools have, despite the limitations of those instruments, proved much more accurate than non-structured clinical assessments.
Static-AS or Automated Sexual Recidivism Scale (ASRS)
[17] The measure commonly used in New Zealand for measuring static risk (risk that remains the same and cannot be changed by the offender) is the Static-AS, now referred to as the Automatic Sexual Recidivism Scale (with the rather unfortunate acronym ASRS). Static-AS or ASRS is a shortened version of the Static-99 measure which was developed in Canada using the technique of meta-analysis. This technique enabled a large number of research studies to be aggregated statistically to gain a more objective assessment of the relationship between predictive variables and the event which is to be predicted. An extensive body of work has been undertaken by both Dr Karl Hanson in Canada and Dr David Thornton in the United Kingdom on the utility and validity of Static-99.
[18] It was necessary to develop a different instrument in New Zealand because three of the ten items in the Static-99 were not routinely captured in the New
Zealand Criminal History Database. These three items are “stranger victim”, “unrelated victim”, and “never lived with a lover for two years”. A programme was developed which enabled the remaining seven variables, which are common to the Static-99 and the New Zealand database, to be electronically scored, forming the basis of ASRS.
[19] A further difference between Static-99 and ASRS is that, under ASRS, prior sentencing dates are counted (see Item 2 in the table at [20] below) and not, as in Canada, charges. This might be seen as more favourable to offenders and could lead to some anomalies depending on the timing of sentencing dates. However, Dr Vess notes that the rationale for using sentencing dates for prior sexual offending is that it shows how many times an individual has been sanctioned and yet persisted in offending.
[20] The risk factors taken into account in ASRS and the weightings attached are set out in the following table:
Static-AS/ASRS Coding Form
Risk Factor Codes Weight
1 Sentences for prior sex offences (exclude current sentence and include only unique result dates)
2 Count of unique prior sentencing dates
(excluding current sentence)
3 Any convictions for non contact sex offences
(all convictions)?
Sentences
None 0
1
2-3 2
4+ 3
3 or less 0
4 or more 1
No 0
Yes 1
4 Current sentence include non-sexual violence? No 0
Yes 1
5 Prior sentences for non-sexual violence No 0
Yes 1
6 Any convictions for male sex victims (all convictions)?
No 0
Yes 1
7 Age as at date coded Aged 25 or older 0
Aged less than 25 1
[21] The scores for each item are then aggregated to find the overall ASRS score. The system currently used by the New Zealand Department of Corrections identifies four risk categories under ASRS, labelled low (ASRS score of 0), medium low
(ASRS score of 1 – 2), medium high (ASRS score of 3 or 4) and high (ASRS score of 5 and above).
[22] Actuarial measures should be validated on the population to which they are to be applied. This has been done with the ASRS. It was applied to all offenders released from prison in New Zealand after serving a sentence for a sexual offence in the years 1987 and 1992 – 2005 (a total of 5,789 cases) and evaluated in terms of predicting sex offender recidivism. In the validation exercise, it was found that the absence of the three items from Static-99 that are not captured in the ASRS did not appear to have reduced the ability of ASRS to discriminate between the risk categories. The development and validation of the ASRS is described in an article by Skelton, Riley, Wales and Vess “Assessing risk for sexual offenders in New Zealand: Development and validation of a computer-scored risk measure” (2006)
12 Journal of Sexual Aggression 277.
[23] ASRS has also been evaluated on specific sub-groups of offenders. The validation of ASRS involved analysis of offending rates at three levels: all sexual offenders and all sexual re-offending, child sexual offenders and all sexual re-offending, and child sex offenders and child sex re-offending. The ASRS is thus empirically validated in relation to re-offending against children by released child sex offenders. The test in s 107I(2) is focussed on the risk of sexual re-offending against children. The relevant figures are thus those for sexual re-offending against children and not sexual re-offending generally.
[24] In assessing re-offending, each re-offender is counted only once (there is no double counting of serial re-offenders). This might mean that the numbers of those who re-offend against children may be understated if a serial re-offender first re-offended against an adult but we still consider it more principled to use the more differentiated figures.
[25] The recidivism rates for released child sex offenders in the various risk categories who re-offend against children are set out in the following table. The percentages in brackets are for general sexual re-offending by child sex offenders.
New Zealand recidivism rates by ASRS risk level for child sex offenders re-offending against children
ASRS Risk Category ASRS Score Percentage of Sexual Recidivism after
5 years
(n = 646)
10 years
(n = 527)
Low 0 2% (2%) 8% (8%)
Medium low 1 – 2 5% (7%) 11% (12%)
Medium high 3 – 4 7% (11%) 16% (24%)
High 5 and above 21% (28%) 36% (43%)
Overall 5% (7%) 11% (13%)
Note: 5 year sample released 1997 and 1998; 10 year sample released 1992 – 1993.
[26] Originally the re-offending rates given to the District Court in this case related to sexual offending generally and not to sexual offending against children. The same appears to have been the case in Barr (see at [8] of that decision). The figures given in Belcher at [80] and [93] for sexual re-offending against children (43% at 10 years) are also for sexual re-offending generally.
[27] Dr Vess pointed out that care must be taken to distinguish a label such as high risk used by the mental health expert and the standards that the Courts will use to determine what preventive measures are warranted under the law. When compared to non-offenders, most sexual offenders will present a high risk of re-offending. In judicial decision making, the concern is more likely to be how an individual offender compares to other offenders. In Dr Vess’ view, the ability for the Courts to assess the relative risk presented by an individual offender can be facilitated by not only providing the rate of re-offending observed in groups of similar offenders but by also assessing the relative number of offenders who populate the various groups. In his view, it is only with this information that users of a risk assessment report will have a general sense of approximately where a particular individual falls in the overall distribution of offenders.
[28] For child sex offenders re-offending against children, the distribution of offenders between the various risk groups is as follows:
Numbers and Percentages of Offenders Within Each Risk Group
Static-AS risk category
Percentage offending after
10 years
Numbers of re-offenders
Percentage of total offenders in each group
Low 8% 19/249 47.25%
Medium-low 11% 23/200 38.00%
Medium-high 16% 08/50 9.50%
High 36% 10/28 5.25%
Overall 11% 60/527
[29] We now turn to the limitations of ASRS. The main issue in using probability statements that are based on observed re-offence rates for groups of offenders (such as those in ASRS) is the degree of inference about the individual that can be made based on findings about a group. Actuarial instruments show the likelihood of re-offending for the group. In Mr Peta’s case, he belongs under the ASRS to a group where 16% of offenders re-offend within ten years. There is nothing in the ASRS that can help assess whether he comes within that 16% or the 84% who do not re-offend. However, this does not mean the ASRS is of no utility. As Dr Vess noted, the fact that an offender belongs to a group in which 16% have sexually re-offended at ten years following release from prison remains a useful measure as it gives a relatively clear understanding of the level of re-offending that has occurred among similar offenders.
[30] The other major limitation of the ASRS is that risk assessments based on static factors are by definition unchanging and insensitive to changes over time in the individual or their circumstances. Many experts recognise that risk is influenced by changing emotional, inter-personal and situational variables, referred to as dynamic risk factors. As noted above at [16], both Mr Riley and Dr Vess agree that this means that the ASRS should be used in conjunction with the assessment of dynamic variables. The Sex Offender Needs Assessment Rating (SONAR) is one measure for
assessing dynamic risk factors. Evaluation work on SONAR has shown a comparatively low correlation with Static-99 scores, suggesting that, while both are able to discriminate between recidivist and non-recidivist sex offenders, they are assessing different features.
Sex Offender Needs Assessment Rating (SONAR)
[31] SONAR grew out of the work initially begun by Dr Karl Hanson in the early
1990s. Contemporaneously, Dr Andrew Harris was involved in research examining the predictors of sex offender recidivism. Hanson and Harris developed SONAR based on data derived from a meta analytical literature review and information obtained from direct field research in Canada and the United States.
[32] SONAR assesses both stable dynamic factors that contribute to risk but change slowly over time and acute factors that may be present for a short period of time and which also contribute to individual sexual offending. The time period assessed for stable risk factors under SONAR is the preceding 12 months. Five stable dynamic factors are assessed:
(a) Intimacy deficits. The scoring criteria for this factor calls, for example, for someone with a current lover and no significant difficulties in that relationship to be scored at zero;
(b)Negative social influences. This calls for an examination of the positive (ie pro-social) and negative (ie anti-social) influences provided by specific relationships in the offender’s life among those who are not paid to be with him. Where more than eight such people are identified, the eight with whom the offender has the most contact are to be considered;
(c) Attitudes tolerant of sexual offending. Particular specified beliefs, which would justify or excuse sexual offending against women or children, are required to be assessed;
(d)Sexual self-regulation. This concerns poorly controlled expressions of sexual impulses and the tendency to use sexuality as a method of coping with negative emotions. It includes attitudes of sexual
entitlement and behaviours consistent with sexual preoccupation. The assessment is based on such behaviours as high levels of pornography use, frequenting strip clubs, massage parlours or prostitutes, deviant sexual fantasies or preoccupation with sexual crimes as well as beliefs that the offender has a higher than usual sex drive, is entitled to sex, or cannot control his sexual behaviour when aroused; and
(e) General self-regulation. The scoring criteria for this factor focuses on the offender’s ability to self-monitor and conform to the demands of community supervision. The areas to assess include, testing known risk factors (such as drinking and drug use), lack of investment in treatment, failing to meet conditions or attend commitments and unwillingness to make sacrifices to avoid high risk situations.
[33] Turning to the scoring of the acute risk factors, what is assessed is whether each of the problem areas has improved, deteriorated, or remained the same during the past month at the time of assessment. For each factor where there has been improvement, one point is subtracted from the SONAR total. For each factor that remains the same, no points are added or subtracted from the total and for each factor that has deteriorated, a point is added to the SONAR total score.
[34] The four acute dynamic factors are: (a) Substance abuse;
(b)Negative mood. This includes feelings of depression, discouragement, anxiety, stress, frustration and loneliness;
(c) Anger; and
(d)Victim access. This includes cruising for sexual victims, attempting to create opportunities to re-offend sexually, the grooming of victims and obtaining a vehicle or engaging in hobbies that would be expected to facilitate contact with potential victims.
[35] Two further measures have been derived from SONAR; the Stable 2000 and the Acute 2000. The Stable 2000 is an enhancement of SONAR and contains six
(rather than five) stable factors. The Acute 2000 expands the four acute factors of
SONAR into eight factors.
[36] Mr Riley explained that current practice when preparing health assessment reports is for Department of Corrections psychologists to undertake a semi-structured clinical assessment by using the factors in the SONAR as anchors to provide a form of checklist of known risk factors. Reference to these factors further informs the assessment of risk, the initial baseline being established by the use of ASRS as a screening instrument. Mr Riley accepts that this means that there is the potential for variation amongst Corrections Department staff in their assessment of the risk factors contained in the SONAR instrument. The Department had, however, only been able to arrange formal training in this measure from August 2000.
[37] Dr Andrew Harris, the co-developer of SONAR, has been retained by the Corrections Department and is facilitating a series of in-depth workshops on the use of this measure. Once all psychological staff have been trained, then the instrument will be “scored” in the formal sense in every case. There is no doubt that this will greatly increase the value of the SONAR assessment. In Dr Vess’ view, once the SONAR is scored properly, no variation between different psychologists would be expected, given the detailed scoring system for SONAR.
[38] Dr Andrew Harris describes SONAR as “empirically informed” rather than “empirically validated”. The amount of data available in published studies remains considerably less than that which supports the Static-99 on which the ASRS is based. There is, however, an increasing amount of data of a statistical nature which indicates that SONAR is a reasonable predictor of sexual re-offending. Currently there is, however, no specific New Zealand validation data in relation to either the SONAR, the Stable 2000 or the Acute 2000. Nor is there direct statistical or numerical translation of SONAR scores into specific re-offence rates.
The revised psychopathy checklist (PCL-R) and sexual deviancy
[39] The existence of psychopathy, particularly if combined with sexual deviancy is a strong indicator of the likelihood of future sexual re-offending. Psychopathy is a
severe form of personality disorder with distinctive emotional, inter-personal and anti-social features. Highly psychopathic offenders are characterised by emotional deficits such as a lack of empathy or remorse, a manipulative and exploitative interpersonal style, and a blatant disregard for the rights of others. Research has consistently found psychopathy to have a strong relationship to a variety of negative criminal justice outcomes. These include poor response to available treatment interventions, increased involvement in institutional misconduct while incarcerated and high levels of violent and sexual re-offending as compared to less psychopathic offenders.
[40] It is generally agreed that the best standardised measure for assessing psychopathy is the revised psychopathy checklist (PCL-R) developed by Dr Robert Hare. It assesses 20 emotional and behavioural features that define psychopathy, based on a thorough review of file and interview information. When it is conducted by adequately trained persons it produces highly reliable scores.
[41] Sexual deviancy is defined as arousal related to sexual activities with children or coercive sex with non-consenting adults. The generally accepted best standardised procedure for assessing deviant sexual arousal is penile plethysmograph technology (PPG). This measures increase in penile circumference in response to various sexual stimuli, usually including visual and/or auditory stimuli. In the absence of such data, the presence of deviant sexual arousal must be inferred from established behavioural patterns associated with deviant sexual behaviour, such as repeated attempts to engage in sexual activities with children or sexual attraction for stimuli representing children.
[42] Hildebrand, de Ruiter and de Vogel examined the sexual recidivism rates among a sample of treated rapists: see “Psychopathy and Sexual Deviance in Treated Rapists: Association with Sexual and Nonsexual Recidivism” (2004) 16 Sexual Abuse: A Journal of Research and Treatment 1. They reported a sexual re-conviction rate of 82% over an average follow-up of 11.8 years for offenders who were both psychopathic and sexually deviant, compared to 18% for offenders who were both non-psychopathic and non-deviant. Similar outcomes have been observed with other samples, including child molesters: see Rice and Harris “Cross-validation
and Extension of the Violence Risk Appraisal Guide for Child Molesters and
Rapists” (1997) 21 Law and Human Behaviour 231.
[43] Mr Riley sounded a note of caution. He said that it must not be assumed that an absence of psychopathy is a protective factor. The absence of psychopathy means that there is one significant risk factor which does not apply but it does not mean that, because an offender is not psychopathic, he or she is safe. In the New Zealand context, the Department of Corrections’ experience at special treatment units for men who offend sexually against children is that the rate of psychopathy in that group of sex offenders is relatively low.
Other relevant factors
[44] Dr Joel Dvoskin and Dr Kirk Heilbrun point out three aspects of risk that should be addressed in the context of decision-making, namely likelihood (ie probability), imminence and severity of outcome: see “Risk Assessment and Release Decision-making: Toward Resolving the Great Debate” (2001) 29 Journal of the American Academy of Psychiatry and the Law 6. Likelihood begins with an assessment of risk based on proven actuarial measures. This is potentially modified by dynamic factors. Imminence of risk is defined by the pattern of violence or sexual violations in the offender’s prior offending in light of their current behaviour, plans, victim availability and life circumstances. Severity is best defined by prior episodes to date including the index offence(s) and the most recent relevant offence(s). Imminence and severity should thus be included as part of the risk assessment.
[45] From the perspective advanced by Dvoskin and Heilbrun, offenders who have demonstrated a pattern of serious offences involving penetrative sexual assaults and the use of abduction or force would be those presenting the highest risk of such serious offences in the future. Conversely, offenders whose offences fall in the less serious range of sexually inappropriate behaviour, if their offending is repeated, would be expected in most cases to demonstrate a level of severity similar to their past offences. This means that the nature of the current and past sexual offending will be a significant factor in any assessment of the s 107F(2)(a) factor which
examines the nature of any likely future sexual offending by the offender: see s 107F(2)(a).
[46] Victim characteristics and previous patterns of sexual offending also bear on the issue of the likelihood of sexual re-offending. There is no doubt that offenders without prior sexual offences re-offend at substantially lower rates than those who have committed prior sexual offences. Child molesters with male victims have consistently shown higher re-offence rates than child molesters with exclusively female victims. Finally, child molesters whose victims are not family members have higher re-offence rates than incest offenders. The gender of previous victims is reflected in the ASRS scoring (see Item 6 in the Table at [20] above), as is prior sexual offending (see Item 1 in the Table at [20] above), but whether victims are family members is not included in ASRS due to the limitations of the New Zealand database. It is included in Static-99: see at [18] above.
[47] A study by Hansen and Harris in 2004 “Sexual Re-offending Rates - Sex Offender Recidivism: A Simple Question” (see Public Safety and Emergency Preparedness Canada’s Website at sppcc.gc.ca/publications/corrections/200403-2_e.asp last accessed 19/02/07) reviewed ten follow-up studies having a combined sample of 4,724 adult male sexual offenders followed for periods up to 15 years. The offender rates per category of offender are set out in the following table:
Subgroup 5 years 10 years 15 years
All sexual offenders 14% 20% 24% Rapists 14% 21% 24% Incest child molesters 6% 9% 13% Girl victim child molesters 9% 13% 16%
Boy victim child molesters 23% 28% 35%
Offenders without previous sexual convictions
Offenders with previous sexual convictions
10% 15% 19%
25% 32% 37%
[48] Finally, the post-release environment must be considered and related to the offender’s history of offending behaviour in any individualised assessment of risk: see Skelton, Riley, Wales and Vess “Assessing Risk for Sexual Offenders in New Zealand: Development and Validation of a Computer-scored Risk Measure” (2006)
12 Journal of Sexual Aggression 277 at 284. Other causative factors such as mental illness will need to be considered. We also note that the professional literature has observed that the more time offenders spend in the community without sexually re offending, the lower their risk of doing so becomes. However, there are as yet no clear findings to quantify the rate of risk reduction.
[49] The factors other than those contained in the actuarial measures that are used to formulate a clinical assessment of risk should be stated explicitly in any health assessor’s report along with their alleged effect.
Best practice in risk assessment
[50] As can be seen from the above review, there are well validated actuarial measures that can help distinguish between higher and lower risk offenders. This is especially the case with those measures that more clearly address the risk presented by specific sub-groups of offenders such as child molesters. The ASRS is one such measure. Findings based on static actuarial measures such as ASRS by definition cannot, however, detect changes in risks over time. Such measures are now augmented by standardised approaches to assessing dynamic risk factors through measures such as SONAR.
[51] The utility of tools such as ASRS and SONAR is only realised when they are properly administered, scored and integrated with other relevant information known to relate to the risk of re-offending. Other known risk factors identified in the research as empirically associated with increased rates of sexual recidivism should be reported on, including sexual deviance and level of psychopathy. The factors, other than those contained in the actuarial measures, used to formulate a clinical assessment of risk and the effect they are said to have must be identified explicitly. Further, it is recognised that risk is contingent on a variety of factors that are difficult or impossible to predict with certainty. Risk assessments should state as clearly as
possible the recognisable contingencies that will influence the degree of risk present. A risk assessment report should also specify as clearly as possible the likely victims and the likely severity of harm of subsequent offences.
[52] Risk assessments and the related judicial decision making for risk management are best informed through an individualised formulation of risk. This should draw upon a variety of different sources of information in an attempt to identify risk factors within an aetiological (causative) framework. This recognises that risk is contingent upon factors that are both environmental and inherent in the individual. Such an approach also helps avoid the shortcomings of a mechanical and potentially formulaic assessment of risk, one that is overly reliant on static historical factors and potentially insensitive to features of the individual that change with time and context. In our view, s 107I(2) in any event requires an individualised assessment.
[53] The results of a properly conducted risk assessment must be effectively communicated to the Court. Adequate training in this is required. When reporting the findings of a risk assessment, comparative categorical labels such as high, moderate or low risk should be qualified by probability statements that give corresponding re-offence rates for groups of similar offenders and the numbers of offenders in each category should be specified: see the tables at [25] and [28] above. Any category or label, such as low, medium or high, should be used consistently in any report.
[54] This summary of best practice relies heavily on Dr Vess’ evidence. Dr Vess’
assessment of best practice was, however, accepted by Mr Riley.
Why we had to deal with the matter afresh
[55] In his decision imposing the ESO on Mr Peta, Judge Dawson set out a description of the offending, summarised the original health assessor’s reports and the submissions and then, under the heading “Decision”, just before imposing the ESO, said the following:
[16] The main consideration the Court must have is that under s 107I of the Act in subsection (1), it states that the purpose of an Extended Supervision Order is to protect members of the community and under subsection (2) the Court may make an Extended Supervision Order if it is satisfied, having considered all matters, that the offender is likely to commit any of the relevant offences referred to in s 107B(2) of the Act. From all of the evidence presented, I am satisfied that the Respondent is “likely to commit any of the relevant offences referred to in s 107B(2)” of the Act. This conclusion is based upon the Health Assessor’s reports which are quite clear that static risks remain in place for at least 10 years. It is accepted that the tests used and research conducted by a psychologist in this area may be in a stage of relative infancy and their use to predict future behaviour may be imprecise. Nevertheless, they are the best tools available at this time and their conclusions should not lightly be dismissed. An Extended Supervision Order will enable the Probation Service to oblige the Respondent to submit to Supervision so as to minimise the risk of future offending.
[56] An ESO has the potential to place major restrictions on the freedom of movement and freedom of association of an offender: see at [12] above. This makes it even more important than in the ordinary course of cases for a Judge, when imposing an ESO, to explain clearly to the offender why such an order is being made. This entails more than a mere reference to the health assessor’s report. More importantly, however, a health assessor’s report should not merely be rubber stamped: see at [7] above. The fact that the reasons given by the Judge were so sparse gives rise to concern that this is what has happened in this case.
[57] A Judge is, of course, perfectly entitled to accept the evidence of a health assessor, particularly in a case where no contrary evidence has been presented. It must, however, be explained why the evidence was accepted and why that leads, on an individualised assessment, to the conclusion that the statutory test for the imposition of an ESO is met. In a case where there is a history of serious sexual offending, where no countervailing factors are identified, where the ASRS and SONAR assessments and any individualised risk factors suggest a high risk of re-offending against children, the Judge’s reasons can be relatively brief. However, in a case like Mr Peta’s which has some unusual features, the analysis should, in our view, have been much more extensive. Mr Peta’s score on the ASRS related to general criminality as against prior sexual offending and there was no evidence of any deviant sexual interest, apart from the indecent assaults and these were effectively attempts only. In addition, Mr Peta’s offending, while clearly unacceptable and distasteful, was at the lower end of this type of offending.
[58] Mr Peta’s sexual offending took place while he was left in the company of a friend’s children for a period of about ten minutes. He was very much under the influence of drugs and alcohol. Mr Peta went into the room where the first victim was sleeping, pulled down his pants, exposed his penis and told her to kiss it. She pretended to be asleep. He then went into another girl’s bedroom. Once there he kissed her on the cheek, exposed himself and asked her to kiss his penis. When the child began to cry Mr Peta left the room. He told the children not to tell their mother and, according to one of the girls in her evidential interview, threatened to kill them. In his written statement Mr Peta admitted being at the house but he claimed not to have done anything at all. He pleaded guilty, however, to the offences and was sentenced to nine months imprisonment.
[59] In addition, the way the Judge’s reasons were framed suggests that he relied solely on Mr Peta’s ASRS score and thus took into account only the static risk factors. As indicated at [52] above, the legislation requires an individualised risk assessment, taking into account all relevant factors. ASRS merely puts an offender into a group of offenders with known recidivism rates but does not predict whether the offender comes (in Mr Peta’s case) within the 16% who re-offend or the 84% who do not: see at [29] above. There was also no specific consideration by the Judge of the factors set out in s 107F(2).
[60] We also consider that the Judge should have specifically addressed the points made by Mr Peta at sentencing. It appears from the sentencing remarks that the points made on behalf of Mr Peta were that the relevant offending was at the bottom end of the scale for sexual offending, that no positive credit had been given to Mr Peta for his genuine on-going relationship with his partner, that in the 18 months since his release from the sentence he had not re-offended and that the tests used could not give an empirical measure of the individual risk posed by Mr Peta to the community. Only the latter was addressed by the Judge in summary form in his reasoning.
[61] It is also necessary for a Judge to consider the term of an ESO in a separate exercise against the statutory criteria in s 107I(5). This was not done by Judge Dawson. In particular, the seriousness of the harm to likely future victims and its
relationship to the other s 107I(5) factors was not considered explicitly. The Judge went immediately from his finding that “static risks remain in place for at least
10 years” (see at [55] above) to the imposition of the ten year term.
The effect of the new evidence on the original health assessor’s report
[62] It turned out, when we heard the further evidence on 30 November 2006, that the problems with what had occurred in the District Court were even more fundamental than appeared to be the case after the July hearing. The Judge had relied on evidence that it is now conceded fell far short of best practice. Indeed, Mr Riley accepted that the original health assessor’s application of the ASRS and the SONAR was not expert in this particular case. It is disturbing that this should be the case. Judges should be able to rely on evidence from the Corrections Department meeting best practice standards. In this case the problems with the original health assessor’s report mean that no reliance can be placed on that report.
[63] The first area of concern with the original report was the mis-scoring of the ASRS test. Ironically, while noting that the test is scored electronically to minimise the level of error, the original health assessor proceeded to score the test manually and wrongly by including convictions in the Youth Court in the category of prior sentencing dates which is against the coding rules. Mr Peta’s risk category on the properly scored ASRS remains medium-high, but the error did make a difference between a score of three and four. While in this case the error did not change the risk category, that would not necessarily be so in other cases. In future, only electronically scored ASRS test results should be presented in evidence.
[64] The second error by the original health assessor was an assertion to the Court that the risk is higher in the second five years for offenders in the medium-high risk category than it is in the first five years (as it appears the Court was told in R v Brown at [53]). This is not correct. The figures are cumulative and thus the ten year rate is bound to be higher than the five year one. For the medium-low and medium-high risk groups, the rates go up by a reasonably steady rate over the ten year period (1% per year for the medium low risk group and 1.5% per year for the medium high group). The risk is therefore constant over the ten year period. For the
high risk group the rate is higher in the first five years (4% per year) than in the second five years (3% per year).
[65] The next error was that the original health assessor treated Mr Peta’s denial of the offending as an indicator of likely recidivism. Research has shown that denial of offending by itself does not operate in this manner. It is true that s 107F(2) of the Act includes the offender’s acceptance of responsibility and remorse of past offending as a factor to be considered. Any denial of offending must be considered, however, in the proper context. While denial has not been found to be not predictive in itself, it may still be a matter that could contribute to an individual’s risk by causing difficulties with willingness to undergo treatment. Thus, while denial is not a predictive factor, recognition of offending can be a protective one. Further, where combined with evidence of psychopathy and sexual deviance, denial of offending can be an important consideration.
[66] The next difficulty with the original health assessor’s report was that, although he purported to provide a SONAR score, he was in fact using Stable 2000 factors. This in itself would not necessarily have invalidated his findings. However, Mr Riley accepted that the assessment conducted by the original health assessor was not a proper assessment in terms of how either the SONAR or the Stable 2000 should be used. The SONAR/Stable 2000 factors were not properly scored but were used by the original health assessor as an aide-mémoire for a semi-structured clinical risk assessment.
[67] We are surprised that it does not appear to have been explained to the District Court Judge in this case (or as far as we are aware in any other case) that the assessment of dynamic risk that was presented in the ESO application was not a SONAR assessment at all but merely a clinical assessment using the SONAR (or in this case Stable 2000) factors. The original health assessor in this case purported in evidence formally to score the SONAR test, although not using the proper structured scoring system. As Dr Vess pointed out, this had the danger of lending pseudo-scientific validity to findings which were not properly based. He said:
In conducting assessments for the Court, failure to follow standardised procedures and recognise scoring criteria for recognised risk measures are potentially worse than not using such measures at all. Erroneous and misleading conclusions may be drawn but appear to have the weight of scientific research behind them, and therefore carry an undeserved weight in legal proceedings where they are introduced into evidence.
[68] Even using the SONAR factors in a semi-structured clinical assessment, the original health assessor did not assess the stable dynamic factors in accordance with the correct SONAR scoring criteria and did not use the acute dynamic factors in the proper manner to reduce or augment the stable dynamic factors. Looking at the stable dynamic factors (see at [32] above), the original health assessor appears to have extrapolated from the sexual offences committed by Mr Peta in this case that he has a belief system justifying sexual offences against children.
[69] Dr Vess’ evidence (which we accept) was that it was unacceptable scoring to infer simply from the existence of an offence that the underlying attitudes supportive of sexual offences must be there. He said that the attitudes and beliefs to be assessed are laid out in the SONAR criteria. Unless the assessor has specifically asked about those or otherwise considered evidence of their presence, it is not possible to conclude whether such attitudes are present or not. Equally, the original health assessor considered that Mr Peta was lacking in sexual self-regulation because of his sexual offences. Again an assessor is required to consider a specific set of indicators under the SONAR and the mere presence of a sexual offence is not adequate to assign the score.
[70] The original health assessment scored Mr Peta adversely with regard to intimacy deficits because of Mr Peta’s alleged hostility to women. He seems to have concluded that Mr Peta is hostile to women on the basis of his convictions for assaults against a previous partner. Again, this was not the correct approach. Dr Vess noted that intimacy deficit factor has more to do with the overall nature of enduring relationships and the capacity for intimacy as opposed to the fact that relationships might be marked by occasional violence.
[71] Looking at the factor of positive and negative social influences, the original health assessor appears to have concentrated merely on an assessment of Mr Peta’s
family and considered that they were a negative social influence. This was in contrast to the assessment of Dr Vess who considered his family to be among Mr Peta’s positive social influences. From our assessment of the files and the various reports on Mr Peta over the years, we agree with Dr Vess’ assessment. We consider that Mr Peta’s family has shown themselves desirous of his living an offence free life. While their efforts to support him in this may not have been successful, the professionals involved also seem to date to have had very limited success.
[72] One of the factors that the original assessor took into account in classifying Mr Peta’s family as a negative social influence was the fact that Mr Peta senior had been convicted of a sexual offence. That sexual offence was committed as a teenager and consisted of consensual sexual intercourse with his then 14 year old girlfriend who is now his wife of 30 years. To suggest this shows an attitude supportive of sexual offending, as the original health assessor insisted it did, has only to be stated to show its unsustainability.
[73] One of the other reasons that the health assessor had for counting Mr Peta’s family as a negative social influence was the fact that the family did not believe that Mr Peta had committed the offences in question. The original health assessor saw this as indicating that the family would not put in place protective measures. He was very influenced by the fact that in Court Mr Peta’s mother had said that she would not hesitate to leave children in Mr Peta’s care.
[74] We have no doubt, however, that Mr Peta’s mother was not contemplating leaving Mr Peta in charge of children when he was in a state of high intoxication through alcohol and drugs, the state he was in when the sexual offending occurred. Dr Vess’ evidence (which again we accept) is that he was satisfied, from talking to the family and to Mr Peta, that Mr Peta and his family would put in place protective measures. He suggested that the mother’s original assertion that she would not do so could well have been related to the context in which the questions were put, ie in the context of her supporting Mr Peta in Court.
[75] In Dr Vess’ view, where there is undeniable and incontrovertible evidence that there is a pattern of sexual offending, a family’s support of an offender’s denial of offending would be worrisome from a risk management and risk prediction point of view. This would be even worse where the family had attitudes not only of denial but supportive of the offending. In this case, there was an isolated incident of offending of a relatively less severe nature which is not accompanied by an attitude that Mr Peta does not need to be mindful of restrictions and risk preventive measures. In Dr Vess’ view, in this case the family’s support of Mr Peta’s denial is not a situation that should cause concern. We agree.
Mr Peta’s ASRS score
[76] Mr Peta’s ASRS score is set out in the following table:
Mr Peta’s ASRS score
Risk Factor Codes Weight Mr Peta’s
Score
1 Sentences for prior sex offences (exclude current sentence and include only unique result dates)
2 Count of unique prior sentencing dates (excluding current sentence)
3 Any convictions for non contact sex offences(all convictions)?
Sentences
None 0 0
1
2-3 2
4+ 3
3 or less 0 0
4 or more 1
No 0 0
Yes 1
4 Current sentence include non-sexual violence? No 0
Yes 1 1
5 Prior sentences for non-sexual violence No 0
Yes 1 1
6 Any convictions for male sex victims (all convictions)
No 0 0
Yes 1
7 Age as at date coded Aged 25 or older 0
Aged less than 25 1 1
Total 3
The effect of the lack of previous convictions and boy victims
[77] Mr Mason, for Mr Peta, submitted that the risk of Mr Peta committing a further relevant offence should be reduced because he had no previous sexual convictions (prior to the index offence) and no convictions involving boy victims. In his submission, the presence of one of these factors more than doubles the risk of re-
offending against a child within ten years. Mr Mason submitted that, in order to assess static risk accurately, Mr Peta should not be included in a group whose recidivism rate is overwhelmingly influenced by offenders with previous sexual convictions and/or boy victims.
[78] Mr Mason relied for this submission on figures provided by the Department of Corrections. These showed that of the 50 child sex offenders released in
1992 - 1993 who were in the medium-high (score 3,4) ASRS risk category, 36/50 (72%) had prior sex sentences (Item 1) and 21/50 (42%) had a conviction for a male sex victim (Item 6). Of the eight child sex offenders in the medium-high category who were re-convicted for sexually offending against children in the ten year period following release, seven out of eight (87.5%) had prior sex sentences and five out of eight (62.5%) had a male sex victim conviction. Of the 50 child sex offenders released in 1992 – 93 in the medium-high ASRS risk category with scores of 3, none had the same contributing items as Mr Peta.
[79] Mr Riley’s evidence was that it is not legitimate from a statistical and mathematical perspective to say that a score of three on ASRS (as Mr Peta had), which is made up of factors that are not related to prior sexual offences or boy victims, is less predictive of re-offending than a score of three made up of those factors. This is because the scales are predicated on the fact that each individual item, such as prior violence and a current violent sentence, contributes uniquely to the degree of risk. The weightings which are assigned to those items in the scale reflect the contribution of that item in the context of the other items in the scale to the predictive accuracy. Had Mr Peta had a boy victim or prior sexual offences then he would have moved from the medium-high into the high risk category. The scale was developed using regression models which allow the attribution to each item of the precise amount of variability in the data which it accounts for. If items are scored equally then they are additive in terms of their predictive effect on the final outcome from a statistical perspective.
[80] Dr Vess agreed with Mr Riley that it is not a useful exercise to deconstruct ASRS but he did not consider it illegitimate to use some of those individual items, such as boy victims or the existence of prior sexual offending, to adjust the risk of
recidivism up or down in the same way as dynamic factors or the existence of other risk factors such as psychopathy are taken into account. We accept this evidence. Equally it seems to us legitimate to make adjustments to the risk assessment (but not to the ASRS score itself) for any apparent anomalies in the scoring system for ASRS, such as appears to have arisen in Barr: see [26] - [28] of that decision.
[81] It is, of course, possible that the actuarial instruments will be further refined. Mr Riley points out that, as increasingly data is electronically scored, there may be ways in which items can be combined by algorithms in the computer that will yield finer discriminations than exist at the moment with an additive scale. For instance, he accepted that some of the relationships between the predictive variables and the outcomes may be non-linear. In the context of an additive numerical scoring system such as used in ASRS, it is impossible to take this into account. Dr Vess also opined that it may well be that the different factors have different effects across different subtypes of offenders (eg those with boy or stranger victims). There is a complex and ambitious research agenda to look at the factors in different combinations. This may lead to further refinement of Static-99 and ASRS.
The alleged double counting of violent offences
[82] Mr Peta’s ASRS result included scores of one in each of the categories “current sentence include non-sexual violence” and “prior sentences for non-sexual violence”: see his ASRS score at [76] above. Mr Peta received a score of one in the former category as his current sentence includes sentences for common assault as well as for doing indecent acts. The incident giving rise to the common assault conviction occurred over ten months after the sexual offences.
[83] Mr Mason submitted that, had Mr Peta been sentenced for the common assault and indecent acts on different days, he would have fallen in the medium-low (rather than medium-high) ASRS category. In Mr Mason’s submission, this fortuitous component of Mr Peta’s ASRS score should be disregarded, leaving his static risk as medium-low. He submitted that, if the Crown wished to maintain the position that Mr Peta’s risk was medium-high, relevant statistics should have been provided to show that a violent offence committed ten months after the sexual
offence but sentenced on the same day makes a unique contribution to the assessment of risk of sexual re-offending. In Mr Mason’s submission, such a proposition is counter-intuitive. The rationale for including current violence must be that non-sexual violence committed in the course of sexual offending increases an offender’s risk of recidivism.
[84] Mr Riley accepted that there is an element of chance as to whether a sentence for non-sexual violence is included as a current offence and that sentencing on a different date can lead to a different characterisation. He noted, however, that, in order for both these items to be included in the risk scale, there had to be a statistically demonstrated relationship between violence in the current sentence and future sexual offending on the one hand and between prior non-sexual violence offences and future sexual offending on the other hand. This would mean that these two items in the risk scale each contribute uniquely to the prediction of further sexual recidivism. We accept this evidence. Further, in our view, there is nothing necessarily counter-intuitive in the proposition that recent violence (as compared to past violence and whether or not it is associated with the sexual offences) is itself relevant to sexual recidivism.
The assessment of Mr Peta performed by Dr Vess
[85] Dr Vess began with Mr Peta’s ASRS medium-high risk category. He then assessed the relevant dynamic factors relating to Mr Peta, applying SONAR. Mr Peta received a score under SONAR that placed him in the category labelled low by the developers of that instrument. Dr Vess adjusted the ASRS level of medium- high risk downwards because of the low score on SONAR. This yielded a medium or moderate level of risk.
[86] The Crown submitted that, in comparing Dr Vess’ and the original health assessor’s assessment of dynamic risk factors, this Court should take into account the differences in Mr Peta’s circumstances during the gap of over a year between the two assessments. Dr Vess said that, if he had assessed Mr Peta in September 2005 when the original assessment occurred, he is doubtful that his assessment would have been very different (except possibly relating to the factor of substance abuse). We accept
his evidence on this point. It is, in any event, irrelevant. This Court must deal with Mr Peta’s situation as it presents now. The original health assessor’s report departed so substantially from the SONAR scoring criteria (see at [68] - [75] above) that we can place no reliance on it.
[87] Dr Vess then tested Mr Peta for both psychopathy, using PCL-R, and deviant sexual arousal. A PPG assessment was not conducted for the latter: see at [41] above for a discussion of PPG. Dr Vess concluded that Mr Peta demonstrates neither psychopathy nor deviant sexual arousal. Mr Peta’s score on the PCL-R was average for offender populations and well below the threshold associated with increased risk for sexual offending. Dr Vess found no patterns of behaviour for Mr Peta which would suggest enduring deviant (specifically paedophilic) sexual interest or arousal.
[88] Moving to an individualised risk assessment for Mr Peta, Dr Vess opined that the main factors that increase Mr Peta’s risk for re-offending include alcohol and drug abuse, lack of employment, excessive unstructured time spent with anti-social and offence-supportive associates, and the presence of potential victims when his judgment and impulse control are diminished by intoxication. Conversely, to the extent that Mr Peta abstains from substance abuse, maintains employment, is in a stable relationship with his partner and avoids anti-social associates, his risk of re- offending will be reduced.
[89] Dr Vess reported Mr Peta as desiring to become a successful provider and parent in his current relationship. Dr Vess considers that Mr Peta requires further development of his maturity and responsibility in order to achieve these goals but that there does appear to be motivation for positive change. It was further noted that Mr Peta had been residing in the community since 31 March 2004 without re-offending sexually.
[90] Turning to the nature of Mr Peta’s sexual offending, Dr Vess considered there to be no empirical or clinical evidence to suggest that any subsequent offending, if it were to occur, would be of substantially more severe nature. This is because of the relative lack of severity of Mr Peta’s sexual offending and the lack of
prior sexual offending and psychopathy. Dr Vess noted several aspects of Mr Peta’s sexual offending that are important in relation to future risks. There was no planning or grooming involved. Once initiated, the offending was not carried through with any great aggression or force. In fact, other than the reported kissing of one girl on the cheek, there was no physical contact involved. This meant that Mr Peta’s specific intent was not strong enough or compelling enough for him to pursue the offending aggressively when the victim declined or became upset and he broke off the attempt very readily.
[91] In Dr Vess’ view, Mr Peta’s sexual offending represents the impulsive acting out of an irresponsible and anti-social person whose judgment and behavioural inhibitions were further diminished by drugs and alcohol. There is no indication of any interest or intent to engage in this sort of sexual activity before or since this occasion. Although there continues to be problematical behaviour such as alcohol and substance use, which is facilitated by ongoing association with peers who may encourage general drug use and general criminal activity, there is no indication that his social network includes anyone who supports or encourages sexual activities with children. Dr Vess’ view is that Mr Peta is most at risk under the influence of substances of committing non-sexual offences.
[92] Dr Vess also noted potential protective factors present for Mr Peta. His family remain supportive of him and desire that he live a lifestyle free of offending. While his mother and partner state that the family does not believe that he committed the sexual offences, they certainly do not support or condone such activities.
[93] Dr Vess also considered that several points in Mr Peta’s developmental history are noteworthy. Up until adolescence Mr Peta’s behaviour appears to have been positive and well adjusted. He is reported to have done well at school and excelled at sport and was a happy and caring boy at home. This contrasts with the pattern displayed by many of the more entrenched and psychopathic sexual offenders whose behaviour is observed to be maladjusted and disruptive from an early age. Mr Peta’s problematical behaviour is consistently described by a variety of sources as stemming directly from his abuse of drugs. However, prior to the current sexual
offences, none of these aggressive and destructive behaviours were of a sexual nature.
[94] Taking into account all of the above factors, Dr Vess assessed the specific risk of Mr Peta committing another sexual offence against a child under the age of
16 to be at the low end of the moderate range. We record that Mr Riley conceded that, on the basis of Dr Vess’ assessment, Mr Peta falls at the lower end of where the Department of Corrections has made applications for the imposition of ESOs.
Is an ESO justified based on the evidence now before the Court?
[95] We turn first to the s 107F(2) factors. Under s 107F(2)(a), we must first consider the nature of any likely future sexual offending by Mr Peta. The seriousness of the offences was relatively low and we accept Dr Vess’ evidence that there is nothing to indicate that any future offending would escalate in seriousness.
[96] Next, s 107F(2)(b) requires the Court to consider the offender’s ability to control his sexual impulses. In our view, it is relevant that Mr Peta’s only convictions for sexual offending are the current offences which, as we have already noted, are towards the lower end of the scale of sexual offending. In this regard, we accept Dr Vess’ evidence that Mr Peta has shown no indication of any interest or intent to engage in this sort of sexual activity before or since. In addition, Mr Peta’s score on the risk assessment tools (both static and dynamic) employed by the health assessor was not based on sexual risk factors. There is no indication of deviant sexual arousal and Mr Peta did not aggressively pursue his offending in the face of distress or resistance from the victims. We do not, therefore, consider that Mr Peta has manifested an inability to control his sexual impulses.
[97] Section 107F(2)(c) requires the consideration of the offender’s predilection and proclivity for sexual offending. This requires a consideration of both actuarial and individualised risk factors. The assessment of static risk under ASRS has Mr Peta in the medium-high category and therefore in the top 15% of offenders, with a group re-offending rate of 16% at ten years. Particular risk factors, such as the existence of boy victims and previous sexual offending, were, however, not present
(although we note that victims were “stranger victims” as against family members which may counteract this to some extent).
[98] We accept Dr Vess’ evidence that the dynamic factors in Mr Peta’s case serve to reduce the risk of sexual recidivism indicated by ASRS. Mr Peta’s dynamic risk is based largely on substance abuse and social influences, although there have been some promising indications over the last year in this regard, particularly in Mr Peta’s wish to provide for his family. Mr Peta does not have a history of committing sexual offences while under the influence of alcohol or drugs and we note that the sentencing Judge accepted that, had Mr Peta not been under the influence of drugs and alcohol, he was not the sort of person who would have offended sexually in that manner. We endorse Dr Vess’ view that Mr Peta’s alcohol consumption places him at a much higher risk for non-sexual than sexual offences.
[99] We also accept Dr Vess’ evidence that Mr Peta has not illustrated attitudes supportive of sexual offending and that his family and associates do not exhibit such attitudes either. We also accept Dr Vess’ evidence that Mr Peta and his family understand the need for him to avoid high risk situations. In addition, Dr Vess found that Mr Peta has not exhibited signs of the significant risk factors of psychopathy or deviant sexual arousal that are associated with increased levels of sexual recidivism.
[100] Under s 107F(2)(d) the Court must also address the offender’s acceptance of responsibility and remorse for past offending. In our view, Mr Peta’s denial of sexual offending is not a cause for concern in the circumstances of this case. As indicated above at [65], denial of offending is not a predictive factor but recognition of offending may be a protective one as it may facilitate acceptance of treatment. Both Mr Riley and Dr Vess recognised, however, that there may be risks with treating sexual offenders at the lower range of seriousness of offending (as Mr Peta is). The treatment of offenders who have committed less serious offences may be counter-productive because of offenders’ exposure to the behaviour of more dangerous and more deviant offenders. In a case like Mr Peta’s, where the denial is not in the face of multiple sources of on-going information about risk, in Dr Vess’ view (which we accept) denial is not a major consideration. Mr Riley agreed that the
priority for treatment in Mr Peta’s case was in relation to his drug and alcohol problems rather than any sex offender programmes.
[101] Further, it is our impression that Mr Peta’s denial of the offending became vehement when the ESO was in contemplation because he did not want to be branded a sex offender guilty of behaviour he recognises as wrong and deviant. Hanson and Morton-Bourgon in “Predictors of Sexual Recidivism: An Updated Meta-Analysis” (2004) (available on Public Safety and Emergency Preparedness Canada’s Website at sppcc.gc.ca/publications/corrections/pdf/200402_e.pdf last accessed 19/02/07) state (at 17) that offenders who minimise their crimes are at least accepting that sexual offending is wrong.
[102] On the basis of his individualised assessment of Mr Peta, taking into account all of the s 107F(2) factors, Dr Vess put Mr Peta’s risk at the low end of the moderate range. For ourselves, we would have put Mr Peta’s risk of sexually re- offending against children as low (but this may be as a result of the difference in perspective between psychologists and Judges noted by Dr Vess: see at [27] above). We are most influenced in our risk assessment by the fact that the sexual offending was not of the most serious character, it was not pursued in the face of resistance, and it appears to have been out of character. In addition, Mr Peta exhibits no signs of a deviant sexual interest in children and has a strong supportive relationship with his partner and family. The most significant risk factor for Mr Peta is alcohol and substance abuse. This is, however, related much more to the risk of non-sexual offending.
[103] We have no doubt that Mr Peta poses a major risk of non-sexual re-offending unless he deals with his substance abuse problem and avoids anti-social associates but it is the risk of Mr Peta sexually re-offending against children that we must assess. On the basis of Dr Vess’ report and our own assessment of the evidence in this case, we are not satisfied that the test in s 107I(2) is met. Mr Peta’s risk of sexual recidivism against children is not such to justify the imposition of an ESO. We are not satisfied that the risk of relevant offending by Mr Peta is both real,
ongoing and one that cannot sensibly be ignored having regard to the nature and gravity of the likely re-offending: see at [8] above.
[104] Had we decided that an ESO was warranted, we would have reduced the term by at least half. Although the level of static risk for those in Mr Peta’s risk category remains constant over a ten year period, Mr Peta’s level of risk as adjusted on an individualised risk assessment is relatively low. The likelihood is that any future sexual offending would be of the same level of (relative lack of) seriousness as the current sexual offending. In our view, a term of ten years would thus have been out of proportion to the nature and risk of any possible future offending.
Result
[105] The appeal is allowed and the ESO imposed on Mr Peta is quashed.
Solicitors:
Crown Law Office, Wellington
44
0
0