New Zealand Law Students Journal
Last Updated: 24 September 2015
PAEDOPHILIC PROCLIVITIES AND THE ASSUMPTION OF HOMOGENEITY
- A PSYCHOLOGICAL CRITIQUE OF THE COURT OF APPEAL JURISPRUDENCE CONCERNING
PROPENSITY EVIDENCE IN CASES OF CHILD SEX ABUSE
At the time this paper was written, a barrage of high-profile paedophile cases assailed media headlines. One such case, dubbed “the worst in New Zealand”, involved Aaron John Ellmers and his sexual abuse of children aged 13 months to 13 years. Further examples include primary school teacher James Parker, confessing to 74 sex crimes against 20 young boys over a 10-year period, and businessman Daniel Taylor, charged with over 20 instances of sexual offending against children. Though shocking, this is not surprising: child sex abuse is not unknown in New Zealand, with the number of incarcerated child sex offenders almost double that of adult sex offenders, and nearly a third of New Zealand women reportedly having been subject to some kind of unwanted sexual contact at a young age.
One of the defining features of child sex is the repetition of sexual violation. Accordingly, ordinary human experience would dictate that in a criminal trial, a jury’s knowledge of the alleged paedophile’s prior sexual misconduct, or “propensity evidence”, would help to ascertain the likelihood of similar offending. In other words, the probative value of the propensity evidence justifies its admission. On the other hand, the sporting theory of justice calls for the exclusion of this evidence in order to prohibit jurors from conveniently generalising the defendant on the basis of adverse traits. Propensity evidence and child sex abuse are a potent mix, one that produces the most severe “knee-jerk” reactions amongst jurors. In a United States Department of Justice survey, the sexual abuse of children, deemed the most vulnerable and innocent in society, was ranked second in heinousness, coming only after the crime of murder. Paradoxically, it is in these emotionally-charged cases of paedophilia that evidence of prior offending is most useful, but also where the reminder of Lord Loreburn LC rings most true: “you cannot convict a man of one crime by proving that he had committed some other crime; that, and all other safeguards of our criminal law, will be jealously guarded”.
New Zealand courts, though, have not paid heed to the Lord Chancellor’s warning. Our courts have allowed decisions of admissibility to be grounded on artificial assumptions which are indefensible with regard to scientific realities. This paper seeks to identify and question these assumptions. It argues that the New Zealand Court of Appeal jurisprudence, since the enactment of the Evidence Act 2006, has wrongly assumed child sex offenders to be a homogenous class when assessing the probative value of propensity evidence, leading to an overestimation of probativeness when comparing similarities in age between the proposed propensity evidence and the trial at hand. The result is an unjustifiable admission of otherwise inadmissible evidence. This paper substantiates its assertion by pointing to recent psychological research demonstrating the heterogeneous nature of child sex offenders, especially in regard to distinct sexual preferences for victim age.
First, in order to establish its argument, this paper will identify the parameters of analysis by defining the concept of child sex abuse. By way of background, Part III will give a brief overview of the relevant propensity provisions in the Evidence Act 2006. Part IV will then focus on how these propensity provisions have been applied by the Court of Appeal in a child sex abuse context. Closer examination of the Court of Appeal’s application reveals an underlying assumption of homogeneity, which will be identified and formulated in Part V. Part VI will then use a psychological framework to critique the assumption identified.
Parameters of Analysis - The Definition of Child Sex Abuse
The term “child sex abuse” is non-exhaustive, and encompasses
various criminal acts of indecency or violation. Legally,
part 7 of the Crimes
Act 1961 is typically invoked in relation to child sex abuse, which includes
charges of sexual violation by
unlawful sexual connection, incest and rape.
Historical abuse may involve now-repealed charges of sodomy or indecency with a
under 12. However, this paper first accepts a more general and
non-exhaustive clinical definition, with child sex abuse being described
“the involvement of adults, older children, or adolescents in sexual activities with children who cannot give appropriate consent and who do not understand the significance of what is happening to them. Such activities violate family and societal taboos. Sexual abuse includes, for example, sexual touching of genitalia, oral sex, attempted or actual sexual intercourse, or including children in child pornography.”
Secondly, for the purposes of this paper, general references to children will be limited to those under the age of 16 to reflect the statutory age of consent to sexual conduct. However, to differentiate between stages of maturity in children aged 0 to 16 years, the Tanner scale of physical development will be adopted. This scale ranges from 1 (prepubertal) to 5 (fully mature), and categorises, using established criteria, sexual maturation with reference to such as breast development in females, genital development in males, and pubic hair growth in both sexes.
Finally, since sexual offending may span over several years or even decades, the age serving as a comparison between various victims of the same offender must be identified. The most appropriate identification is the age the defendant allegedly commenced the abuse against the victim, as this manifests the earliest age at which the offender was prepared to abuse victims. Not only have clinical studies used this method of classification, but courts have likewise done so when comparing similarities in ages between two sets of offending. In R v L(CA276/2009), the abuse against a prior complainant occurred over one year (5 to 6 years of age). The abuse at trial, however, lasted for five years (5 to 10 years of age). The Court of Appeal nonetheless held the defendant’s propensity to be a tendency to engage in “sexual behaviour with a girl of about five.” This method of classification will be essential in later discussions concerning how courts have erroneously framed defendants’ propensities.
The Propensity Provisions - Section 40 and 43 of the Evidence Act 2006
Prior to the enactment of the Evidence Act 2006, the common law doctrine of
“similar fact” evidence allowed admission,
in a criminal trial, of
any evidence probative of disreputable aspects of the defendant’s
character. Application of the
doctrine involved a balancing of competing interests: a weighing of the
probative value of the proposed evidence
against the unfair prejudice the
evidence may have on the fact-finder’s
reasoning. The Evidence Act has
largely retained the common law analysis relating to propensity
However, the Supreme Court in Mahomed v R confirmed that any inquiry
regarding interpretation of the propensity provisions must “focus firmly
on the terms of the
A “clean state” approach, with little to no reliance being placed on
pre-Act case law, is therefore
Where the Crown wishes to adduce propensity evidence against the defendant, the evidence must first qualify as “propensity evidence” within s 40(1) of the Act. Section 40(1) acts as a mechanism to define the types of evidence to which the following propensity provisions apply. Assuming the proposed evidence falls within s 40(1), the probative-prejudice calculus in s 43 is engaged. Where, having regard to enumerated discretionary factors, the probative value of the evidence outweighs the risk of unfair prejudice to the defendant, the propensity evidence is then admissible.
A. Section 40(1): Definition of “Propensity” Evidence
Propensity evidence is evidence “that tends to show a person’s propensity to act in a particular way or to have a particular state of mind”. Concepts of linkage and coincidence form the rationale of propensity evidence. The Supreme Court in Mahomed recognised that the greater the linkage between the propensity evidence and the present charge, the greater the probative value of the propensity evidence. Nonetheless, the linkage identified must not only have some sort of specificity, but that specificity, though must also be connected to the conduct or mental state alleged to constitute the litigated charge. The threshold for specificity, though, in acts or conduct is not extremely high, and courts have even found a propensity to have no relevant convictions.
B. Section 43: The Probative-Prejudice Calculus
1. Section 43(1): the overarching test
Assuming the evidence is defined as “propensity” under s 40(1) and is offered by the prosecution, the s 43(1) test is engaged to determine admissibility of the propensity evidence, which is possible
“...only if the evidence has a probative value in relation to an issue in dispute in the proceeding which outweighs the risk that the evidence may have an unfairly prejudicial effect on the defendant.”
The statutory test therefore aims to address the tension inherent in propensity evidence by balancing two competing interests: that of probative value to a material issue in the proceeding, and that of unfair prejudice to the defendant.
2. Section 43(2): framing the relevant issue
Section 43(2) makes clear that the nature of the issue in dispute must first be identified. How that material issue is framed largely impacts the assessment of probative value, and in turn whether the evidence should be admitted at all. The 3-2 Supreme Court split in Mahomed, a case concerning parents charged with murder of their child, stemmed from a divergence in the definition of the material issue. Whereas the majority confined the issue to being one of intention for murder,  the minority argued that the issue was one of “hostility” and “indifference towards the child’s exposure to risk.” It must be noted that in cases of child sex abuse, however, there are rarely disputes as to the material issues: since consent is not a valid defence, most cases concern the defendant’s identity or whether the conduct in question actually took place.
3. Section 43(3): factors in determining probativeness
The first step of the s 43(1) balancing test is a s 43(3) inquiry into the probative value of the propensity evidence, which is the tendency of any evidence to prove or disprove a material issue in the litigated trial. In doing so, the court may have regard to the factors outlined in s 43(3). Though these factors are not a “checklist” concluding with “some sort of arithmetical ‘totting up’ of what is present against what is not”, the majority of judgments have methodically worked through these factors when determining probative value. These factors, to paraphrase, are as follows (emphasis added):
(a) Frequency of offending;
(b) Connection in time between proposed evidence and present charge;
(c) Extent of similarity between the proposed evidence and present charge;
(d) Number of persons making allegations against the defendant;
(e) Whether such allegations in (d) may stem from collusion or suggestibility;
(f) Extent to which the acts are unusual.
Section 43(3)(a) indicates that the more acts done which demonstrate the specific propensity (frequency of offending), the higher the probative value of the propensity evidence. Evidence that a defendant committed rape on 10 prior occasions is clearly stronger evidence of a propensity than if a defendant had merely committed one rape. Section 43(3)(b) indicates that probative value is heightened the closer in time the prior conduct had been committed to the litigated charge.
Section 43(3)(c) has been central to judicial analyses of probativeness. Similarities in the prior acts and the litigated charge are to be identified, and the greater the extent of similarity between the two, the more probative the propensity evidence. However, the jurisprudence relating to paragraph (c) has been criticised, both judicially and academically. Commentary has argued that there is an over-emphasis on similarities, often to the exclusion of blatant differences. This is particularly true in cases of child sex abuse, and forms the basis of the current assumption of homogeneity argued in this paper.
Sections 43(3)(d) and (e) concern an examination of the complainants. According to paragraph (d), the probativeness of propensity evidence increases with the number of complaints. However, paragraph (e) warns against the possibility of collusion between those complainants, and against the potential for particularly suggestible complainants (for example, children). Finally, section 43(3)(f) notes that the sharing of unusual features between the two sets of offending may serve to increase probative value.
4. Section 43(4): Considerations of unfair prejudicial effect on the defendant
Considerations of probativeness are fundamentally connected with notions of fairness. Unfair prejudice may be where the jury risks adopting an emotional, “knee-jerk” reaction to evidence as opposed to rationally applying it to the facts. Accordingly, the determination of the propensity evidence’s probativeness under s 43(3) must be balanced against any illegitimate prejudice to the defendant (s 43(4)). As noted by Baragwanath J in R v B (CA459/06), s 43(4) signifies Parliament’s recognition that jurors are human, and human nature is essentially guided by simplistic reasoning, such as “no smoke without fire”, or “he’s a bad man, he’s done it before”. Sections 43(3)(a) and (b) therefore compel considerations of the risk of an unfair predisposition against the defendant, or possibilities that the jury may accord disproportionate weight to the propensity evidence. Thus, for any evidence defined as “propensity evidence” under s 40(1) to be admitted, it must pass the high threshold balancing test in s 43(1), being deemed more probative to a material issue than unfairly prejudicial to the defendant.
An Overview of Court of Appeal Jurisprudence: Propensity Evidence in Cases of Child Sex Abuse
Child sex abuse proceedings have almost invariably required judicial consideration of propensity evidence and its admissibility. Guidance from the highest court is scarce: the only Supreme Court decision concerning s 43 and child sex abuse involved consideration of whether evidence presented in trials culminating in an acquittal could be adduced as propensity evidence. A unanimous Court held that even prior acquittal evidence may be adduced if it satisfies the requirements of s 43(1) of the Act.
On the other hand, the New Zealand Court of Appeal has grappled extensively with the issue of admission or exclusion of propensity evidence relating to alleged paedophiles. The s 43 balancing test has been invoked in the context of severance applications, appeals against conviction, or pre-trial interlocutory proceedings relating to admissibility. An examination of the Court of Appeal’s position when applying propensity evidence to circumstances of child sex abuse circumstances will subsequently reveal an assumption of homogeneity.
A. The Definition of a Paedophilic Propensity: Section 40(1)
The Court of Appeal has generally defined a propensity for paedophilia by reference to victims in the two sets of offending. Propensities, though, have ranged in specificity. Where the proposed evidence and the present charge are more alike, the Court frames the propensity as a specific one. This is highlighted in the pre-trial ruling of Rhodes v R, which concerned the admissibility of previous relevant convictions where similar sexual activity involving vaginal and anal penetration was carried out. The defendant was held to have the specific propensity to “engage in diverse forms of sexual activity with young persons... [particularly inserting] objects in the vagina and/or anus of a young person in his care.”
Alternatively, propensities have been defined in general terms where the proposed evidence and the litigated charges are different, the Court endeavouring to admit the evidence without straining its propensity definition, as in Lindsay v R. The defendant was appealing against conviction on counts of unlawful sexual connection and attempted sexual violation against two adolescent girls. The propensity evidence proposed by the Crown concerned a sexual grooming conviction. The majority, in admitting the propensity evidence, framed the defendant’s propensity generally as a tendency to “act upon his sexual interest in young girls in his care.” Mackenzie J dissented, criticising the lack of specificity in the majority’s formulation of the defendant’s propensity. He argued that having a broad “sexual interest in young girls” does not cover the minor offence of grooming and allegations of outright sexual violation, being actions which are significantly different. 
B. Determining the Probativeness of Prior Instances of Sex Abuse: Section 43(3)
Admission of propensity evidence hinges upon its probative weight. Like in other non-sexual cases, courts, in determining probativeness, have applied the factors enumerated in s 43(3) to proceedings involving child sex abuse.
1. Frequency and connection in time: paragraphs (a) and (b)
An analysis of the frequency of offending under s 43(3)(a) in child sex abuse proceedings has usually uncovered prolonged and sustained abuse by the defendant. In O v R for example, the Court of Appeal identified the alleged offending as occurring almost weekly over eleven and three years for the first and second complainants respectively, and over three years for the propensity witness. These were clearly not “isolated incidents”. Alternatively, the s 43(3)(b) factor of connection in time is discussed mostly in historic offending cases, where probative value is typically weakened by a wide gap in time between the two sets of offending. In Lowe v R,  the Court of Appeal held that a gap of at least 32 years between the present charge and the prior evidence lessened its probative value.
Nonetheless, the probative value of propensity evidence under paragraph (b) comes down to judicial selection of comparable events. The Court of Appeal O v R preferred to place more emphasis on the time elapsed between the first and second complainants, and the first complainant and the propensity witness, but not the second complainant and the propensity witness. In doing so, it avoided connection in time issues which would have amounted to a gap of at least 39 years.
2. Extent of similarity: paragraph (c)
The s 43(3)(c) factor concerning extent of similarity undoubtedly engenders the most judicial debate. This is unsurprising: the very nature of propensity evidence is based on coincidence and linkage principles, involving an identification of similar aspects present in both the allegation and past conduct, which ultimately feeds into formulation of the defendant’s propensity. However, Court of Appeal jurisprudence is unsettled as to whether analysis of similarities necessarily includes analysis of comparable differences. Traditionally, the Court has focused exclusively on similarities, thereby admitting incidents of prior child sex abuse even in the face of notable dissimilarities. The Court in Rhodes, for example, first acknowledges that a charge of rape and unlawful sexual connection against two 6-year-olds is quite different to that of inserting a vibrator or beads into the anus/vagina of a propensity witness. However, it then comments that in such contexts, “the emphasis is not so much on differences between separate events...as on similarities between them”, going even further to say that “an approach that focuses on differences rather than similarities runs the risk of being unduly narrow.”
The approach of Glazebrook J in Vuletich v R, a rape case against two female adult complainants, may be contrasted to Rhodes above. In Vuletich, “obvious differences” between the two sets of offending, including the different degrees of consent (equivocal consent in the first offending, as opposed to outright disgust in the second) outweighed the similarities, ultimately rendering the propensity evidence inadmissible. Her Honour’s approach is supported by commentators, who deem that analysis of the “extent of differences” is inherent in the paragraph (c) analysis.
The reasoning of Glazebrook J has been applied in the context of child sex abuse. In S(CA567/2011) v R,  an opportunistic offending against a 19-year-old male was not cross-admissible as propensity evidence in other offending against three other males, who were 13 to 16 years of age. The Court, in severing the complaint brought by the 19 year old, recognised that there were indeed similarities, especially concerning the nature of the offending (oral and anal sex) and the presence of alcohol. However, it stated that “the ‘extent’ of similarities can only be fully appreciated by an assessment of their strengths and weaknesses”, and that “material differences...cannot be ignored in [the] assessment process.” Therefore, at least in some cases of child sex abuse, differences between the two sets of offending are identified to ensure the evidence is not unfairly admitted.
3. Number of complainants and the potential for collusion or suggestibility: sections 43(3)(d) and (e)
The majority of cases which have come before the Court of Appeal have involved multiple complainants who were unaware of the existence of one another. However, even where complainants were previously acquainted with one other or in some way connected, that was not enough, as the allegations of child sex abuse had to be the result of collusion or suggestibility.
4. The unusualness of the propensity evidence/present charge: section 43(3)(f)
Generally, the Court has found child sex abuse to be inherently unusual: “the interest of a mature male in sexual activity with very young girls is, itself, an unusual factor”. However, it is not sufficient for merely one of the alleged propensity incidents to be unusual, as the unusual features “must be common to both sets of offending for this factor [of unusualness] to be met.” Therefore, in R v S(CA567/2011), the performance of oral sex and sodomy by an older man against younger men in one set of offending was unusual, but oral sex and adult homosexual sex in the other set was not inherently unusual. Since both sets of offending did not share the same unusual features, this factor was not met.
C. Concerns of Unfair Prejudice to the Alleged Paedophile: Section 43(4)
While evidence of prior sexual abuse may be highly probative, the propensity evidence must still be excluded if there is risk of an unfairly prejudicial effect on the defendant. Juror simulation studies have proved, in relation to prior paedophilic conduct, that “more is worse, more is believable”. The Court has also recognised this special and inherent prejudice experienced by defendants in trials involving child sex abuse, since “the existence of previous convictions might of itself lead the jury to conclude ‘once a paedophile, always a paedophile’.” In most cases, prejudice may be overcome by a judicial direction to the jury. However, in other cases, the Court has found that unfair prejudicial effect would remain notwithstanding any direction given by the trial judge. This is particularly where allegations involve historic sex abuse, or where the propensity evidence adduced led to an acquittal in a previous trial.
Overall, s 43(1) considerations of probativeness and prejudice in child sex abuse proceedings are inextricably tied. Where probativeness is slight, it would clearly be unfairly prejudicial to the defendant to admit the propensity evidence. As such, correct evaluation of the probativeness of prior sexual abuse, free from any invalid assumptions, is essential if the defendant is to be protected under the 2006 Act.
Identifying an Assumption of Homogeneity
A. The Assumption of Homogeneity in the Court of Appeal Jurisprudence
A critical examination of the current approach outlined above reveals a misunderstanding regarding the behaviour of paedophiles. The Court of Appeal overestimates the probative value of propensity evidence involving paedophilia by relying on the assumption that paedophiles constitute a homogenous class of offenders, whereas in reality, there is no “classic personality pattern for p[a]edophilia because of the various sub-groups that exist.” The assumption of homogeneity is therefore the judicial misunderstanding that child sex offenders do not discriminate between pre-pubescent and pubescent victims. This assumption goes further than the above criticisms of Vuletich and S(CA567/2011) v R. It does not centre on the fact, as noted by Glazebrook J, that judges are failing to consider significant differences under paragraph (c). It instead highlights that, in cases of child sex abuse, the very process by which those similarities and differences are assessed is grounded on an assumption unjustifiable in light of psychological evidence. Judicial failure to recognise this distinction has resulted in wrongful admission of inaccurately weighted evidence.
Nonetheless, before the assumption can be critiqued, it must first be exposed. The assumption of homogeneity can be seen clearly, simply by comparing the ages of child complainants in the two or more sets of offending in cases mentioned above. The Court of Appeal in R v Shaw framed the defendant’s propensity as a tendency to indecently assault “young girls”, a tendency which “most obviously” justified five charges of indecent assault being heard by the jury against five different complainants. In that case, the youngest of these complainants was 6 years old, whereas the eldest was 12 years old.  The Court of Appeal equates the youngest child as similar to the eldest, and effectively assumes that the defendant would have the propensity to sexually abuse all children in that age range. Similarly, in the severance application of O v R, the Court of Appeal accepted the trial judge’s observations regarding the ages of the child victims, in that they were “young” at the time of offending.  This was even though their ages ranged from 7 years to 12 years of age.
Lindsay v R concerned a prior grooming conviction involving an 8-year-old being adduced in the trial of two 14-year-old complainants.  The fact that the majority allowed the propensity shows their understanding that having a sexual interest in “young girls” meant having one that encompassed the ages of 8 to 14. The same type of reasoning is evidence in Rhodes, where the Court of Appeal allowed the admission of propensity evidence involving an 11-year-old in a trial involving two 6-year-old claimants. The Court noted that it would not “mechanically identifying points of similarity and dissimilarity” between the two sets of offending. By reasoning that considerations of the “extent of similarity” under s 43(3)(c) should not encompass considerations of differences, the Court of Appeal effectively side-stepped the complication regarding the obvious difference in age.
Various other cases not noted above also show this underlying assumption. The defendant in Rompa was charged with indecent assault, inducing an indecent act, and sodomy on a 5-year-old male.  The Crown sought to adduce propensity evidence of a female witness alleging that the defendant had sexually abused her when she was 10 years old. While the Court did highlight the discrepancies between the age and sex of the two complainants, one being a 5-year-old male and the other being a 10-year-old female, it did not attach much weight to these differences, and promptly dismissed the appeal without further discussion regarding its admission. The Court of Appeal adopted the propensity formulated by the trial judge, which was to engage in “conduct relating to young children, whether male or female”.
Finally, in R v Davidson, which involved a severance application on representative charges of unlawful sexual connection and indecent assault, the Court recognised that there were “some differences” in the offending. This included an “age difference” between the two sets of offending. However, such a difference was ignored by the Court of Appeal, which deemed both female complainants were “pre-adolescent”, and therefore sufficiently similar. It must be noted that the propensity victim was a child of 5 years, whereas the trial at hand concerned a 12-year-old. Like the five other examples identified above, Davidson demonstrates the assumption of the Court of Appeal that the alleged paedophiles have a sexual interest in a broad range of victims, regardless of age.
The fact that the Court of Appeal permits evidence of a 6-year-old to be tried on the same indictment as a 12-year-old, as in Shaw, or when the Court allows a prior conviction involving an 8-year-old to be adduced into a trial brought by two 14-year-olds, highlights an underlying conception that child sex offenders have a uniform sexual preference for all children under the age of 16. Alleged paedophiles are therefore assumed to sexually violate victims indiscriminately, with no regard for differences in age and varying degrees of physical or sexual maturity. This is the assumption of homogeneity.
B. Can the Current Approach be Rationalised?
While the judicial assumption that all paedophiles share sexual preferences for a broad range of child victims may be disproved by clinical research, there may nonetheless be reasons why courts have admitted such dissimilar evidence.
1. Rationalisation #1: the unique nature of child sex abuse proceedings
Judges are tasked with viewing disputed evidence in a global manner to account for the implications of admissibility on the entire trial. Therefore, the special characteristics of child sex abuse would inform and shape the final decision on admissibility. First, child sex abuse occurs in familial contexts or positions of trust, frequently taking place in secluded locations. In Lowe for example, the indecent assault, whereby the defendant allegedly placed his hand down the front of the complainant’s jeans, took place in the private and isolated setting of the defendant’s bedroom. As such, independent witnesses are unavailable to testify in favour of the complainant’s allegations, meaning the only directly incriminating evidence is that of the victim’s testimony, which the defendant would invariably deny. Like in many sexual violation trials, proceedings then essentially devolve into assessments of credibility, as to whether the child complainant or the defendant is more believable.
Another unique feature to child sex offences is that child victims are typically plied with treats or money to conceal the sexual nature of the relationship, or are forced to stay silent and refrain from reporting the offending by verbal or physical threats. When the sexual abuse is ultimately disclosed, sometimes years or decades after the alleged offence, there is a lack of medical evidence otherwise corroborating the victim’s account of abuse. This undue delay in bringing proceedings, caused by fear of threats or by a failure to comprehend the nature of the indecent touching, then paradoxically becomes a factor weighing against the victim. Thus, in Lowe, historical acts of offending occurring between 1976 and 1978, more than 40 years prior to the trial, led to the inability of prior complainants to correctly recollect the time and circumstances of the original abuse. The Court held that this risked unfair prejudice to the defendant and accordingly the propensity evidence was excluded to the detriment of the complainants.
These factors, at times collectively appearing in an individual case, render it difficult for the prosecution to prove beyond reasonable doubt its case against the defendant. Therefore, when faced with cases involving a disparate gap between the ages of the prior victim and the complainant at trial, the Court may find it justified to decide in favour of the Prosecution. This is simply to redress issues of trial fairness, and to ensure that juries receive the benefit of the propensity evidence in order to gain a more complete picture of the defendant.
2. Rationalisation #2: propensity evidence as necessarily broad to be flexible and fact-specific
The current assumption of homogeneity may also be rationalised by arguing that propensity evidence should not be unduly constrained by attempts to create a blanket, “consistent” approach. Child sex abuse often involves abuse of a multitude of victims in a variety of circumstances, which demonstrates the need for a flexible and fact-specific approach under s 43. Distinguishing propensity evidence and present charges on the basis of victim age could thus be viewed as inappropriate quibbling over minor details. McGrath and William Young JJ, when criticising the Mahomed majority, warned against an “atomistic” analysis of individual events focusing on “substantial discrete and independent incremental probative value in the associated evidence.”
However, while flexibility is paramount to a fair trial and underlies the fundamental role of the judge as gatekeeper for evidence, wrongful assumptions must be corrected to ensure that the Court’s “discretionary” analysis under s 43 is exercised in accordance with psychological realities. This is the case for the judicial misconception that paedophilic offenders are a homogenous category. Thus, the many occasions where the Court of Appeal has explicitly conceded the differences in victim ages, but ultimately dismissed those differences, stems not merely from a wrongly-held assumption regarding the compulsive behaviours of paedophiles, but also from judicial cognisance of the unique character of child sex abuse proceedings and the fundamental necessity of flexibility. Nonetheless, although the assumption of homogeneity may be rationalised, it cannot remain tenable when viewed in light of empirical research.
A Psychological Critique of the Assumption of Homogeneity
A. Paedophilia vs. Hebephilia: Hebephilia as Distinct from Paedophilia
1. Historical recognition of hebephilia
Social sciences literature has repeatedly underscored the heterogeneous nature of child sex offenders. Heterogeneity in paedophilia includes categorisations and subdivisions according to age, sexual orientation and typological distinctions. Clinical phallometric assessments employing penile plethysmography, where subjects are interviewed regarding their sexual history then presented with audio-visual stimuli to measure arousal states, have substantiated this heterogeneity. For example, a study published in 1986 comparing erection responses to slide and audiotaped stimuli showed delineation between homosexual and heterosexual paedophiles, incest cases, and a miscellaneous diagnostic group, measured with regard to the differing degrees of arousal. As a result of the extant literature, convenient generalisation by therapists and law enforcement agents of paedophiles as consisting of a single category has been berated as displaying “ignorance” and drawing “premature conclusions” regarding the characteristics of child sex abuse.
While there are many subcategories of child sex offenders, the distinction which has been most problematic for the Court of Appeal has been that relating to victim age (and the corresponding physiological development). Examination of the jurisprudence above shows appellate judges to perceive paedophiles as predators attracted to “children”, whether infants or adolescents. The prevalence of this misplaced perception may, however, be linked to the relative novelty of the delineation between offenders who sexually prefer pre-pubescent children and those who prefer pubescent children. Victim age as a factor to discriminate amongst the general class of paedophiles did not emerge until the 1950s, where psychiatrist Bernard Glueck studied and categorised 200 sex offenders based on their offence history: incest, rape, paedophile and hebephilia. Glueck defined “hebephilia” as “sexual activity with adolescents...as distinguished from paedophilia, which we limit to children below puberty”. Later, Czech sexologist Kurt Freund, the developer of the penile plethysmograph, invoked Glueck’s terminology of “hebephilia” as describing the heterosexual attraction of paedophiles to adolescents. It was also Freund who defined the “hebephilia” as the erotic preference for pubescent victims, corresponding to the age bracket of 11 to 14 for girls and up to 16 for boys.
2. Blanchard’s definitive study on hebephilia
Professor Ray Blanchard, incumbent director of the Centre for Addiction and Mental Health in Toronto, conducted a study, published in 2009, validating the existence of hebephilia as a distinct preference. The study consisted of a large-scale examination of 881 patients over a decade, and involved analysis of the correlation between subjective self-reports and objective psychophysiologically-assessed erotic responses, the latter evaluated by volumetric plethysmography. This is a phallometric procedure measuring penile blood volume change when exposing subjects to audio and visual stimuli. Blanchard’s research was to discern whether “men who report maximum sexual attraction to pubescent children have greater penile responses, in the laboratory, to depictions of pubescent children than to depictions of younger [persons]?”
Methodologically, stimuli to which a particular subject was exposed distinguished explicitly between pre-pubescent and pubescent narratives according to the Tanner classification outlined above. The particular age of the photographic models was made explicit to the subject in the audio stimuli, while also being manifested in their physical maturity (breast, genital, and pubic hair development proportional to their classification on the Tanner stages). Fictional pre-pubescent children were defined as being 5 to 9 years of age (Tanner stage 1) whereas pubescent children were 11 to 13 years (Tanner stage 3). This is identical to the parameters adopted by Freund in his 1981 studies on hebephilia.
The results of the study indicated the overwhelming majority of subjects had attractiveness ratings profiles (gathered from the surveys) matching the age of the group to which they were assigned (discerned from the volumetric plethysmography indexes). Comparisons of hebephiles within their respective groups demonstrated that greater penile responses resulted when exposed to depictions of pubescent children as opposed to depictions of younger or older persons. Between-group comparisons distinguished between men who reported maximum sexual interest to pre-pubescent children from those who reported maximum sexual interest to adults. Although the same trend was not observed for homosexual hebephiles, this was attributed to this category’s smaller sample size, which lessened statistical power. Blanchard observed that the study “showed that hebephilia exists, and – incidentally – that it is relatively common compared with other forms of erotic interest in children”, and therefore recommended that the definition of “paedophilia” in the Diagnostic and Statistical Manual V be expanded to include erotic attraction to pubescent children Whereas he notes there may be divergences in physical maturation and onset of puberty between pubescent girls and boys, using the age range of 11 to 14 years as the definition for hebephilic preference would not “lose much precision”.
3. Other studies substantiating the existence of hebephilia
(a) Other biological assessments
Blanchard and other researchers have also conducted various biological assessments proving the existence of hebephilia as a distinct disorder to paedophilia. Paedophiles and hebephiles have been discriminated in terms of IQ, where those with greater phallometric responses to younger, pre-pubescent children (paedophiles) corresponded to a lower IQ than those with a greater response to pubescent children (hebephiles) or adults (teleiophiles). Other measures separated the two groups by reference to differences in completed education and stature. Comparisons using these factors have shown that hebephiles differ from average population controls, being intermediate in the academic qualifications and in stature relative to paedophiles and teleiophiles.
(b) Implicit Association Test
Implicit measurement techniques have also demonstrated the existence of hebephilia. Such techniques are based on a “response interference paradigm”, which involves subjects rapidly sorting images of models according to age and sexual preference. The associations between child versus adult and sex versus non sex are then examined. Brown and others, in an Implicit Association Test published in 2009, hypothesised that paedophilic would show “distinct patterns of implicit associations between children and sex”. In that study, subjects were required to classify quickly and accurately as possible non-erotic pictures as either adult or child, and words as either sex or non-sex related. “D-scores” would be generated, where scores less than zero demonstrated a greater association between child and sex, while scores greater than zero indicated a greater association between adult and sex. The results of the study confirmed the “large” differences hypothesised between paedophiles and hebephiles, with paedophiles showing an implicit association between concepts of children and sex (D-scores for paedophiles were negative), and hebephiles failing to do so (D-scores for hebephiles were positive).
(c) Analyses of criminal and personal history
Literature examining the criminal and personal histories of paedophiles and hebephiles has also attempted to prove the difference between paedophiles and hebephiles, especially when ascertaining the most appropriate treatment options for the two.  Generally, paedophiles are older than other sexual offenders, and engage in less nonsexual criminal conduct. They are often described as lacking in self-esteem, highly defensive, and possess a background of dysfunctional relationships. However, they have a specific erotic preference for pre-pubertal children and feel the need to please the victim as opposed to just pleasing themselves, which would lead to a high rate of recidivism against similar victims. Hebephiles, on the other hand, are likely to have age-appropriate relationships, a stable family life, and often even having children at the same age of their victims. Causes of offending are unlikely to stem from true sexual preference, but may be caused by external stressors. Since the characteristics of offenders are reflective of and congruent with the developmental period of their victims, treatment options for the two classes should be distinguished accordingly.
B. Criticisms of the Paedophilia-Hebephilia
1. Critics of Blanchard’s findings
The above studies show, therefore, that Blanchard’s recognition of hebephilia is not a novel finding, nor is it unsupported by other empirical studies. However, the clinical studies undertaken by Blanchard and his team were ultimately a means to achieve the goal of re-defining paedophilia in the DSM-V to account for hebephilic offenders. This has not been without controversy, and indeed the thrust of critical literature has revolved around the notion that, by defining hebephilia as a mental disorder, psychiatry is wrongly pathologising something which is commonplace and even normal.
Critics have argued that sexual preference for adolescent children does not necessarily imply the existence of a mental disorder, simply because both the physiology of a pubescent and adult individual is relatively similar. The only reason why sexual conduct with 14-year-olds is condemned by society is due to the artificiality when determining the legal age of consent to sexual intercourse – for example, the Spanish age of consent is 13 years. Accordingly, the arbitrary manner in which consent has been criminalised consent should not be the basis upon which a novel mental disorder is founded: hebephilia is not a “dysfunction”, and therefore should not be classified as a mental disorder. The corollary to this argument is that difficulties may be experienced by clinicians when attempting to apply an objective age at which sexual attraction is considered “normal”.
2. Responding to Blanchard’s critics
While criticisms of Blanchard’s study may be well-founded in a psychiatric context, their applicability is limited in a legal context. First, the legal age of consent in New Zealand is 16 years. A higher level of imprisonment is imposable if the child is under 12 years at the time of the offending. The answer to the first criticism, therefore, is that at least in New Zealand, any discussion on propensity evidence is necessarily constrained by the relevant legislative provisions. Judges are obliged to apply this threshold, even if deemed arbitrary or contrary to what is considered appropriate sexual conduct by psychiatrists: finding the existence of a mental disorder and criminalisation of sexual conduct are vastly different inquiries.
Secondly, in terms of the corollary argument, potential psychiatric difficulties in demarcating the upper range of adolescents and adults also do not translate into legal difficulties. Using the age of consent as a reference point, courts have correctly recognised, for example, that propensity evidence involving a 19-year-old male is inadmissible in a severance case concerning three other defendants aged 12, 13 and 16 years. Likewise, in R v Taua, the Court of Appeal deemed propensity evidence of a 19-year-old inadmissible in the case against a 12-year old. Thus, psychiatric difficulties in drawing the line at which hebephilia ceases to become abnormal does not apply to courts, which are capable of drawing at least general distinctions between child sex abuse and non-child sex abuse.
Finally, such criticisms do not negate the proposition crucial to this paper’s argument. This paper does not purport to argue that hebephilia is a distinct mental disorder or that it should be integrated into the DSM-V. It simply asserts that there is a fundamental difference between a preference for pre-pubescent and pubescent children for the purposes of propensity evidence. This is enough if proved empirically, which has been done by Blanchard and other researchers. Clearly, a difference in sexual preference for a propensity analysis does not need to amount to a psychiatric disorder, before it can be termed a “difference”. The identification by Blanchard’s critics of peripheral issues relating to the hurdles associated with a new clinical diagnosis of hebephilia does not, therefore, detract from the essential finding that hebephilia is a distinct sexual preference and is of high relevance to the determination of probative weight in propensity evidence.
C. Application of the Paedophile-Hebephile Distinction to the Court of Appeal Jurisprudence
Recent empirical evidence has demonstrated that hebephilia, defined as a sexual preference for pubescent children, male or female, between the ages of 11 to 14 years, is distinct from paedophilia, involving sexual preference for pre-pubescent children. Since hebephilia and paedophilia denote different sexual preferences, it is evident that the current jurisprudence of the Court of Appeal, outlined above, fails to account to accord with psychological reality and developments. However, to reinforce this argument, Blanchard’s findings may be explicitly applied, by way of example, to the three cases of Shaw, Lindsay, and Davidson.
In the severance case of R v Shaw, the youngest complainant, a 6-year-old, was to be tried in the same case as the eldest, a 12-year-old. This was because the Court of Appeal deemed the evidence of both sets of offending to be cross-admissible against the other. The defendant in Shaw was assumed by the court to prefer both 6-year-olds and 12-year-olds. An application of Blanchard’s findings to this case would mean that the Court of Appeal wrongly assigned a higher probative value when determining the cross-admissibility of evidence: the eldest complainant there would comfortably be defined as pubescent, whereas the youngest is clearly a pre-pubescent child.
The defendant in Shaw would not be both a paedophile and a hebephile, as they constitute different sexual preferences. Thus, the fact that the defendant abused the 6-year-old complainant at Tanner stage 1 of sexual maturity cannot tend to prove that the same defendant abused the 12-year-old complainant, presumably categorised as Tanner stage 3. While it is difficult to speculate the weight placed on this similarity in age by the Court of Appeal when determining probativeness, it is clear that the Court viewed it as trite and undeserving of attention: “most obviously, all the alleged offending was against young girls.” Blanchard’s findings therefore challenge this banality and dismissive attitude of the Court when assuming homogenous characteristics of child sex offenders, epitomised here as the defendant in Shaw.
The case of Lindsay v R may also be critiqued in a similar manner. The majority’s finding that the propensity evidence involving a 9-year-old was admissible allowed the jury to consider that evidence in a trial concerning pubescent girls aged 14 years. The prior complainant, at 9 years of age, would still be considered pre-pubescent, whereas the older female complainants would surely have begun to show signs of sexual maturity. The majority’s justification of the difference once again displayed the assumption of homogeneity: it amalgamated paedophilia and hebephilia, two distinct sexual preferences, as a single propensity, and then ascribed it to the defendant. Quoting Solicitor-General v Rudd that the “key similarity is the sexual misconduct between an adult man and a young girl,” the Lindsay majority applied Rudd to the facts and held that “all were young girls in [the defendant’s] care as a trusted caregiver”.
Mackenzie J, the minority in Lindsay, dissented and argued that the acts were not sufficiently similar to display a propensity to sexually abuse young girls. However, his Honour, like the majority, still assumed that child sex offenders prefer both pre-pubescent and pubescent children: “any similarity arises only because of the...age of the victims.” If Blanchard’s findings were to be taken into account when deciding Lindsay, the Court of Appeal would have had difficulty reconciling the two sets of offending, and perhaps would have reconsidered the probative value assigned to the propensity evidence. Further, given Mackenzie J’s strong dissent regarding the other obvious dissimilarity relating to the nature of the two charges, which involved a contrast between sexual grooming and sexual violation offences, the Court of Appeal should have given less weight to probative value in light of the differences in the two charges and the ages of the complainants. This is because instead of serving as evidence tending to show the defendant’s propensity to sexually abuse young girls, the evidence of the prior 8-year-old complainant actually militates against the finding of a propensity, assuming Blanchard’s findings of heterogeneity are accepted.
Finally, in the severance application of R v Davidson, the evidence of a 5-year-old complainant was deemed cross-admissible against another complainant aged 12 years (and vice versa). Again, the Court of Appeal, when engaging in the s 43 analysis, wrongly based its finding on the assumption of homogeneity, calling both children “pre-adolescent”. In effect, the Court deemed that all offences against children aged between five and twelve years were sufficiently similar, though one of the victims was clearly a pre-pubescent child, and the other a pubescent one. Accordingly, the physical differences pertaining to sexual maturity between the two victims meant that the defendant would likely prefer one over the other, and is actually evidence which does not tend to show the defendant sexually abused one or the other. The Court of Appeal should have allowed the appeal and severed the charge so that the two counts could be assessed separately, given the lack of probativeness and the heightened risk of unfair prejudice on the defendant. This judicial assumption in Davidson (and other cases) occasioning unfair prejudice clearly shows the need for solutions to be found. Awareness must be drawn to this oft-cited but quickly demonised area of law.
The fundamental tenet of the common law is the presumption of innocence. Detractions from this presumption must be justified. The current statutory test, under the Evidence Act 2006 for admitting propensity evidence against the defendant by the Crown demonstrates an awareness of this need for justification. To avoid trial by character, s 43(1) dictates that the probative value of the propensity evidence must outweigh the risk of unfair prejudice. Although the risk of inflaming the jury is inherent in the admission of any prior misconduct, such a risk is compounded when abuse involving the innocence of society’s most vulnerable victims is concerned. Due to this overwhelming risk of unfair prejudice, any assumptions upon which a decision of admissibility is based must be valid and truly reflect the actual propensity of the defendant.
This paper has argued that the Court of Appeal holds an invalid assumption regarding the homogeneity of offenders, an assumption which does not accurately reflect the true propensity of paedophile defendants and ultimately leads to artificial exaggeration of probative value. It has examined the jurisprudence relating to the admissibility of propensity evidence in cases of child sex abuse, pointing out in the jurisprudence an underlying judicial assumption that pre-pubescent and pubescent children herald the same attractiveness for sex offenders. It has shown that this assumption is untenable in light of recent psychological research, which demonstrates that child sex offenders have distinct sexual preferences according to victim age.
Propensity evidence in general is an area attracting surprisingly sparse literature. Propensity evidence in cases of child sex abuse attracts even fewer commentators still. This is despite the vast amount of litigation revolving around s 43 of the Evidence Act 2006 and the frequent media publications involving paedophilia. Nonetheless, this does not relieve the New Zealand judiciary from failing to take into account evidence, this time of a scientific nature, crucial to the liberty of New Zealand citizens faced with allegations of paedophilia.
 William D Murphy and others “Differential Sexual Responding Among Four Groups of Sexual Offenders Against Children” (1986) 8 Journal of Psychopathology and Behavioral Assessment 339 at 350–352.
 At 339–340.
 At 346–347.
 At 348.
 At 172.
Bruce Rind and Richard Yuill “Hebephilia as Mental Disorder? A Historical, Cross-Cultural, Sociological, Cross-species, Non-Clinical Empirical, and Evolutionary Review” (2012) 41 Archives of Sexual Behaviour 797 at 803.