The High Court has dismissed an appeal against a decision of the New South Wales Court of Criminal Appeal on tendency evidence in the context of multiple child sexual offences. The appellant, a well-known actor in a 1980s television series, was convicted of nine child sexual offences and sentenced to 10 years and nine months imprisonment. Among the evidence at trial was evidence from a range of complainants and other witnesses on the appellant’s sexual interactions with them, which was said to establish a tendency of the appellant to act in a particular way or have a particular state of mind, specifically, holding a sexual interest in children, using his social, familial and employment relationships to gain access to them, and engaging in particular kinds of sexual conduct. The NSWCCA dismissed his appeal against the conviction and sentence, rejecting (among a number of other arguments) that the trial judge erred in allowing the tendency evidence on the basis that their statements were dissimilar and related to dissimilar circumstances to the acts charged, and thus lacked significant probative value as required by s 97 of the Evidence Act 1995 (Cth). Section 97 provides that evidence adduced for the purpose of proving that an accused had a tendency to act in a specific way, or have a specific state of mind, will be inadmissible unless the court ‘thinks that the evidence will, either by itself or having regard to other evidence adduced or to be adduced by the party seeking to adduce the evidence, have significant probative value’.

Before the High Court the appellant sought to contend that the NSWCCA erred in its application of the law on tendency evidence, specifically in finding that the tendency evidence here had significant probative value given the dissimilarities in the nature, context and circumstances it described, and that the NSWCCA further erred in not requiring some ‘underlying unity’ or ‘pattern of conduct’, and in rejecting the approach taken by the Victorian Court of Appeal in Velkoski v The Queen [2014] VSCA 121 of assessing degrees of similarity in considering possible probative value.

The High Court, by 4:3 majority, dismissed the appeal. The majority (Kiefel CJ, Bell, Keane and Edelman JJ) held that ‘in a case in which the complainant’s evidence of the conduct the subject of the charge is in issue, proof of that tendency may have that capacity.’ (at [2]).

Dealing with the second ground first, the majority reviewed the ALRC interim report on tendency evidence (see [20]ff), and the cases considered by the NSWCCA (at [25]ff), before considering Velkoski (at [29]). In the majority’s view, while the VSCA correctly observed that common law principles on similar fact evidence had been superseded by pt 3.6 of the Evidence Act, the majority was critical of the Court of Appeal’s reliance on common law language — particularly ideas of ‘underlying unity’, ‘pattern of conduct’ and ‘modus operandi’ and the criterion of ‘similarity’ protecting against the risk of an unfair trial (see at [31]) — which ‘do not stand’ with the new scheme in pt 3.6 (at [32]). The VSCA’s approach glossed the language of s 97, and the absence of any reference to similarity or the familiar common law terms just mentioned indicates that the section is not to be read as if it contained those ideas or terms (at [34]).

Contrary to the VSCA’s approach in Velkoski of assessing the probative value of tendency evidence by looking to similarity of ‘operative features’ of the acts proving the tendency, the majority held that ‘a tendency to act in a particular way may be identified with sufficient particularity to have significant probative value notwithstanding the absence of similarity in the acts which evidence it’ (at [37], and illustrating this approach with reference to the facts in Velkoski at [38]). Ultimately, the probative value will depend on the factual issue the tendency evidence is being used to prove, and different considerations may be important where the fact in issue is the occurrence of the offence itself (at [39]). The majority took up the articulation of s 97’s test from the earlier case of Ford: that ‘the disputed evidence should make more likely, to a significant extent, the facts that make up the elements of the offence charged’, noting only that it is not necessary that the disputed evidence has this effect ‘by itself‘ (at [40], emphasis in original).

That assessment in turn requires first, considering the extent to which the evidence supports the tendency, and, secondly, ‘the extent to which the tendency makes more likely the facts making up the charged offence’: at [41]:

here the question is not one of the identity of a known offender but is instead a question concerning whether the offence was committed, it is important to consider both matters. By seeing that there are two matters involved it is easier to appreciate the dangers in focusing on single labels such as “underlying unity”, “pattern of conduct” or “modus operandi”. In summary, there is likely to be a high degree of probative value where (i) the evidence, by itself or together with other evidence, strongly supports proof of a tendency, and (ii) the tendency strongly supports the proof of a fact that makes up the offence charged.

Unlike the common law tests and principles it replaced, s 97 has a more ‘open-textured’ inquiry and evaluative task, which, despite likely difficulties and differences of views among courts in marginal cases, requires the ‘application of the same well-known principles of logic and human experience’ also used to assess the relevance of evidence: at [42].

Returning to the first ground (on whether the tendency evidence had significant probative value) the majority held that it did have that capacity (at [57]):

the tendency evidence showed that the unusual interactions which the appellant was alleged to have pursued involved courting a substantial risk of discovery by friends, family members, workmates or even casual passers-by. This level of disinhibited disregard of the risk of discovery by other adults is even more unusual as a matter of ordinary human experience. The evidence might not be described as involving a pattern of conduct or modus operandi – for the reason that each alleged offence involved a high degree of opportunism; but to accept that that is so is not to accept that the evidence does no more than prove a disposition to commit crimes of the kind in question.

Here, the force of this tendency evidence was not any indication of repeat offending, but rather that, in this accused’s case, the complaint should not be rejected as being ‘unworthy of belief’ merely because it seemed improbable, in that it was outside of ordinary conduct: at [60]. Applying the two inquiries mentioned at [41], the majority held that the evidence, when considered together, ‘provided strong support to show the appellant’s tendency to engage opportunistically in sexual activity with underage girls despite a high risk of detection’ (at [62]), and that this tendency made more likely the elements of the offence charged (at [64]). Consequently, the NSWCCA did not err in concluding that the tendency evidence met the requirements of s 97: at [64].

Gageler J, Nettle J and Gordon J each dissented.

After noting the many barriers to admissions of tendency-type evidence (at [68]ff) Gageler J emphasised that the tendency rule is aimed at addressing a form of cognitive bias common in inferential or inductive reasoning; that is, ascribing disproportionate weight to past conduct as an indicator of future conduct, and treating a person as ‘more consistent than he or she actually is’ (see at [70]–[73]). Gageler J then noted that this understanding is consistent with the legislative history on the rule, and particularly the ALRC interim report (see at [74]ff). Gageler J stated that the degree to which tendency evidence might rationally affect the assessment of the probability of the existence of a fact depends on two considerations: the extent to which the evidence (alone or with other evidence) could rationally affect the assessment of the probability that a person, first, has a tendency to act in a certain way or have a certain state of mind; and secondly, having acted or had a state of mind on an occasion in issue in the proceeding (see at [90] and [93], with a number of illustrations from [94]ff).

Turning to Velkoski, Gageler J considered the VSCA’s two statements on similarity, ‘underlying unity’, ‘pattern of conduct’ and ‘modus operandi’ (specifics at [103]) to be ‘unobjectionable’: the VSCA view on similarity captures the purposive approach to applying the tendency rule, and the unity, conduct and modus operandi phrases describe the ‘normal process of tendency reasoning’ rather than laying down any exhaustive test for admissibility: at [104]. As applied to this case, while the tendency notice was specific and ‘elaborate’ in its description, it effectively amounted to a tendency to have a sexual interest in female children and the use of social, family or professional networks to pursue that interest (at [108]). The abnormality of any interest all of this kind makes it difficult to use experience as a guide to evaluating the tendency (at [109]):

A tendency to have such a sexual interest and to engage in sexual activities with female children less than 16 years of age, opportunistically or at all, is so abnormal as to allow it to be said that a man shown to have such a tendency is a man who is more likely than other men to have engaged in a particular sexual activity with a particular female child on a particular occasion. Yet the problem is this: how much more likely is not easy to tell, in part because common experience provides no sure guide, and the abhorrence any normal person naturally feels for such a tendency highlights the risk that any subjective estimation of the likelihood will be greater than is objectively warranted.

Gageler J was unsatisfied, on the material available, that the general tendencies of either sexual interest in or opportunistic sexual activities with female children bears on the probability of specific allegations ‘to an extent that can be properly evaluated as significant’, and that it was not open for the NSWCCA ‘to conclude on the basis of its capacity to establish such a tendency that the evidence of all of the complainants met the tendency rule so as to be cross-admissible.’: at [111] (and see the specific analysis at [112]ff). Because the NSW DPP accepted that if the evidence of any complainant was not cross-admissible against any other complainant that an error of law would have occurred, Gageler J would have quashed the convictions and ordered a new trial (see at [115]–[116]).

Nettle J held that the tendency evidence in relation to each count was wrongly admitted. Nettle J proceeded by taking an ‘orthodox approach’ to the application of s 97: although s 97 clearly intended to state the principles freshly, and its language determines the way in which tendency evidence must be treated, Nettle J held that ‘logically and necessarily’ s 97 involved ‘the same process of probability reasoning that was applied at common law’ (at [174]ff). For Nettle J, probative value required more than the mere ‘unusualness’ of the kind of offence (which his Honour noted was not unusual) or even the fact of an offence having been committed, for a tendency to be established (at [157]. [158]). Something more is needed, such as a ‘logically significant’: degree of similarity in relationship of accused to complainant; connection between details of each offence; a modus operandi or system of offending; ‘or, otherwise, some logically significant underlying unity or commonality, howsoever described’: at [158]. Similarly, ‘opportunism’ by itself seems too broadly expressed to fit this ‘something more’ requirement, and will depend on the details: at [159].

Nettle J also rejected the NSWCCA’s process of reasoning in concluding there were two tendencies (sexual interest generally combined with engaging in sexual conduct in the linked contexts of social, familial and professional relationships): the second of the tendencies ‘amounted to dividing up the general tendency said to be relevant to all counts … into the constituent elements of that tendency and then treating each constituent element as if it were a separate tendency of significant probative value in relation to all counts. It was not’: at [160]. Nettle J emphasised that what makes a modus operandi significantly probative is that the fact that an accused was ‘alleged to have committed each offence in the same or a substantially similar way made it significantly more probable that each complainant was telling the truth’, and even where a similar modus operandi is used that would not, of itself, make it significantly more likely that another child was telling the truth about an offence committed in different circumstances or using a different modus operandi: at [164].

Turning to the evidence here, Nettle J held that this problem of difference had arisen. The counts relating to sexual offences committed at homes could establish a tendency to take advantage of custody, authority or control over female children within private homes (at [165]). The counts relating to the appellant’s daughter’s friend likewise involved custody, authority or control, but the nature and circumstances of the offending was significantly different and there did not seem to be sufficient similarity or unity about that evidence to conclude that its probative value outweighed its prejudicial effect (at [166]). The final counts, relating to sexual offences committed within a reciprocated relationship and in the workplace were in very different circumstances than those of the earlier counts, and the evidence adduced in relation to those counts did not significantly increase the probability of the appellant having committed the other offences (at [167]). Consequently, the trial judge and NSWCCA erred in holding that the evidence was cross-admissible throughout the counts. Having reached that conclusion, Nettle J then turned to reasons not to depart from the ‘orthodox’ approach, examining the pre-s 97 case law (at [175]ff) and its legislative background (at [182]), rejecting that there was any justification for ‘lowering the bar’ of admissibility in relation to unity requirements (at [194]ff), and concluded with comments on the arguments around Velkoski (at [204]ff).

Gordon J ‘substantially’ agreed with Nettle J’s reasons on the main points of probative value requirements here and articulating eight principles underlying that approach (see at [216]). Her Honour added that these principles reflect three matters: the dangers of tendency evidence; the ‘conservative approach’ that is necessary in requiring ‘more’ (such as a ‘logically significant underlying unity or commonality’); and that broadly termed tendencies without such a similarity, unity or connection would run contrary to the way s 97 operates and was intended to operate (at [217]–[219]). That final matter was clearly illustrated by the tendency evidence of the wardrobe assistants relating to count 11: the NSWCCA did not explain how the work environment evidence related to the two ‘essential tendencies’, which shows the difficulties of relying on highly generalised tendencies and of not undertaking the analysis separately and in relation to each piece of evidence (at [220]–[224]). Gordon J concluded that the inadmissibility of the complainant in count 10 in relation to the other counts meant, on the proscutor’s concession, that the remaining counts be quashed and a new trial ordered.