Hostname: page-component-745bb68f8f-v2bm5 Total loading time: 0 Render date: 2025-01-08T16:57:20.090Z Has data issue: false hasContentIssue false
  • English
  • Français

Simply Unconvincing: The High Court on Probative Value and Reliability in the Uniform Evidence Law

Published online by Cambridge University Press:  01 January 2025

Jason M Chin Open the ORCID record for Jason M Chin [Opens in a new window] ,
Gary Edmond Open the ORCID record for Gary Edmond [Opens in a new window]  and
Andrew Roberts
Jason M Chin*
Affiliation:
School of Law, University of Sydney, Sydney, NSW, Australia
Gary Edmond
Affiliation:
Faculty of Law and Justice, University of New South Wales, Sydney, NSW, Australia
Andrew Roberts
Affiliation:
Melbourne Law School, The University of Melbourne, Melbourne, VIC, Australia
*
* The author may be contacted at jason.chin@sydney.edu.au or jmichaelchin@gmail.com
Get access Rights & Permissions [Opens in a new window]

Abstract

Exclusion of evidence when its probative value is exceeded by its risk of creating unfair prejudice has long been a fundamental safeguard against unfair trials and wrongful convictions. In 2016, IMM v The Queen (IMM) curtailed that safeguard by holding that trial judges should assess probative value on the assumption that the evidence is reliable and credible. The IMM majority placed emphasis on the capacity of the evidence. In doing so, it provided a mysterious qualification: some evidence may lack probative value not because it is unreliable, but because it is ‘simply unconvincing’. The majority illustrated unconvincingness with the example of an unreliable eyewitness identification. Courts and legal scholars criticised the majority judgment for its harmful implications and for its apparent incoherence. From a review of almost 4 years of post-IMM jurisprudence and deeper exploration into one particular case, we find that ‘simply unconvincing’ has accentuated the confusion and inconsistency in Australian evidence jurisprudence.

Type
Articles
Information
Federal Law Review , Volume 50 , Issue 1 , March 2022 , pp. 104 - 127
Copyright
Copyright © 2022 The Author(s)

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1. Tasmania v Farhat [2017] TASSC 66 [41] (‘Farhat’) (emphasis added).

2. R v Morris [1983] 2 SCR 190; Federal Rules of Evidence, 28 USC r 403; Police and Criminal Evidence Act 1984, c 60, s 78; Pentland v The Queen [2020] QSCPR 10 [53] (‘Pentland’).

3. Uniform Evidence Law (‘UEL’). The UEL is incorporated into the laws of the Commonwealth (Evidence Act 1995 (Cth)), the Australian Capital Territory (Evidence Act 2011 (ACT)), New South Wales (Evidence Act 1995 (NSW)), the Northern Territory (Evidence (National Uniform Legislation) Act 2011 (NT)) (‘NT Evidence Act’), Tasmania (Evidence Act 2001 (Tas)) and Victoria (Evidence Act 2008 (Vic)), as well as Norfolk Island (Evidence Act 2004 (NI)).

4. S 137 reads: ‘In a criminal proceeding, the court must refuse to admit evidence adduced by the prosecutor if its probative value is outweighed by the danger of unfair prejudice to the defendant’.

5. [2016] HCA 14 (‘IMM’).

6. Ibid [44], [47], [50]. This position deviates from other accusatorial trial jurisdictions, including Australian states that have retained the common law: Pentland (n 2) [53]; White Burgess Langille Inman v Abbott and Haliburton [2015] 2 SCR 182 [24].

7. ‘On the present state of the law, therefore, the judge in a criminal trial is unable to perform the “gatekeeper” Role…’: Chris Maxwell, ‘Preventing Miscarriages of Justice: The Reliability of Forensic Evidence and the Role of the Trial Judge as Gatekeeper’ (2019) 93(8) Australian Law Journal 642, 643. See also Gary Edmond, ‘Icarus and the Evidence Act: Section 137, Probative Value and Taking Forensic Science Evidence “At Its Highest”’ (2017) 41(1) Melbourne University Law Review 106; David Hamer, ‘The Unstable Province of Jury Fact-Finding: Evidence Exclusion, Probative Value and Judicial Restraint after IMM v the Queen’ (2017) 41(2) Melbourne University Law Review 689; Andrew Roberts, ‘Probative Value, Reliability and Rationality’ in Andrew Roberts and Jeremy Gans (eds), Critical Perspectives on the Uniform Evidence Law (Federation Press, 2017) 63. Two practitioners expressed confusion over it: Stephen Odgers and Richard Lancaster, ‘The probative value of evidence’ (Winter 2016) Bar News 36.

8. See, eg, ss 97, 97A, 98, 101, 135, 137.

9. Benjamin Johnson and Richard Jordan, ‘Should Like Cases Be Decided Alike?: A Formal Analysis of Four Theories of Justice’ (2018) <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3127737>.

10. IMM (n 5) [114] (Nettle and Godon JJ), contrasting reliability with credibility (or the whether the witness is intentionally being truthful or not).

11. One possible manner is through rhetoric, part of the classical training of lawyers in centuries past. See, eg, the political speeches or defence work of Cicero. We doubt, however, that the High Court is concerned with the art of persuasion.

12. This sample of cases was assembled as part of a larger project aimed at advancing transparent and reproducible empirical legal analysis, with these cases described in a searchable app that researchers and practitioners can review and use to address different questions. See Jason M Chin, ‘An open and reproducible analysis of a controversial criminal law decision: Mission IMM-possible?’ (2020) <https://osf.io/preprints/lawarxiv/2j3ye>.

13. Courts are more or less experienced with different types of evidence. Although, this essay reveals that courts may be quite limited in their understanding of the limits of some types of evidence, such as voice recordings, which appear frequently before them, see Part IV.

14. R v Ellis [2003] NSWCCA 319 [95] (Spigelman CJ) (‘Ellis’).

15. IMM (n 5) [29]; R v Shamouil (2006) 66 NSWLR 228 (‘Shamouil’); R v XY (2013) 84 NSWLR 363 (‘XY’).

16. IMM (n 5) [23], [44], [47], [50]; Edmond (n 7) 119.

17. Dupas v The Queen (2012) 40 VR 182 (‘Dupas’).

18. The alleged conduct began when she was 12: IMM (n 5) [1] (French CJ, Kiefel, Bell and Keane JJ).

19. Ibid [3], [15]-[16]. For instance, s 97 of the Evidence (National Uniform Legislation) Act 2011 (NT) requires that tendency evidence (eg evidence of the defendant’s previous similar deeds) have ‘significant probative value. And s 101 requires a balancing of that probative value against its prejudicial effect.

20. NT Evidence Act (n 3).

21. Ibid s55(1).

22. Ibid Dictionary pt 1 (defintion of ‘probative value’).

23. IMM (n 5) [39].

24. The majority proceeded to explain that in a very narrow range of cases, evidence will be so ‘inherently incredible, fanciful or preposterous’ as to be capable of rationally affected assessment of facts in issue, and as a consequence, would not satisfy the requirement of relevance: Ibid [39]. Such a finding appears to be have been seldom made: Chin (n 12) 34.

25. IMM (n 5) [39].

26. Ibid [49].

27. Ibid [44].

28. Ibid [50].

29. IMM (n 5) [50].

30. This was originally provided by Heydon, Hon JD Heydon, ‘Is the Weight of Evidence Material to Its Admissibility?’ (2014) 26 Current Issues in Criminal Justice 219, 234.

31. Ibid (emphasis added).

32. See generally, Wayne A Davis, Meaning, Expression, Thought, (Cambridge University Press, 2000). Many contemporary theories of meaning draw on the work of Grice: HP Grice, ‘Meaning’ (1957) 66 The Philosophical Review 377; HP Grice, ‘Utterer’s Meaning and Intention’ (1969) 78 The Philosophical Review 147.

33. David Lewis, ‘Languages and Language’ in Keith Gunderson (ed), Minnesota Studies in the Philosophy of Science, Volume VII, (University of Minnesota Press, 1975) 3, 15.

34. Ibid.

35. Note that reliability has a scientific meaning, but lawyers tend to treat it as trustworthiness or robustness of the evidence.

36. See, eg, Daubert v Merrell Dow Pharmaceuticals Inc, 509 US 579 (1993), 113 S Ct 2786; R v Trochym, 2007 SCC 6, [2007] 1 SCR 239; Law Commission of England and Wales, Expert Evidence in Criminal Proceedings in England and Wales (The Stationery Office, 2011).

37. References to ‘unconvincing’ do appear in the NSW line of cases that the majority preferred. It was used once in Shamouil (n 15) [42] to explain why the exclusion of an eyewitness identification at trial weakened the Crown’s case enough that it could appeal the decision. The court in Shamouil found that the identification was important because although the witness later retracted it, a jury could ‘take the view that the attempt to retract the identification evidence was unconvincing’. After establishing this was a proper appeal point, the court held that the identification should have been admitted. The New South Wales Court of Criminal Appeal in XY (n 15) [62] later quoted Shamouil’s use of ‘unconvincing’ to say that the retraction of the identification was a matter for the jury and should not impact the identification’s probative value. But it had not been used, as it was in IMM, to describe a quality of evidence.

38. See Alena Skalon, Mehera San Roque and Jennifer L Beaudry, ‘An Interdisciplinary and Cross-national Analysis of Legal Safeguards for Eyewitness Evidence’ in M Miller and B Bornstein (eds), Advances in Psychology and Law (Springer, 2020) 137; United States, National Research Council, Committee on Science, Technology, and Law, Identifying the Culprit: Assessing Eyewitness Identification (National Academies Press, 2014).

39. Jury Directions Act 2015 (Vic) s 36; Domican v The Queen (1992) 173 CLR 555.

40. Indeed, the majority draws an explicit distinction between reliability and unconvincingness, but provides no explanation for that distinction, IMM (n 5) [50]: ‘…on one approach it is possible to say that taken at its highest it is as high as any other identification, and then look for particular weaknesses in the evidence (which would include reliability). On another approach, it is an identification, but a weak one because it is simply unconvincing […] it is the latter approach which the statute requires’.

41. The parameters of this database can be found in a paper focused on its development: Chin (n 12).

42. We broadened the search to the word ‘unconvincing’ rather than ‘simply unconvincing’ to account for the chance that some courts might not use the entire phrase. We also broadened the scope to pre-IMM cases to see if simply unconvincing had been explored before IMM appeared to introduce it.

43. ‘I am unable to be satisfied that that evidence has significant probative value on that issue. It is a “simply unconvincing” case, to use the language of IMM…’: R v Al-Harazi [2016] ACTSC 250 [31]. See also Tasmania v C [2017] TASSC 9 [20]-[22].

44. R v Eastman (32) [2018] ACTSC 12 [37]-[64], [85]-[97] (‘Eastman 32’);

45. Bembo v The Queen [2018] VSCA 42 [33]-[34], [40] (Whelan JA).

46. Eastman 32 (n 44).

47. Ibid [37]-[39].

48. Ibid [95].

49. Ibid [96]; here it cited R v Debresay [2016] VSC 487 (‘Debresay’), which we will see offered of the most thorough analyses of the foggy night example.

50. Eastman (32) (n 44) [96].

51. Other decisions have purported to apply ‘simply unconvincing’ without explaining why the evidence did or did not meet this threshold. This has occurred with potentially concocted and contaminated statements: BM v The Queen [2017] NSWCCA 253; statements containing incorrect dates: R v Jenkin (No 10) [2018] NSWSC 705; children’s memory for events: A2 v The Queen [2018] NSWCCA 174; and unconscious displacement: R v Eastman (No 31) [2018] ACTSC 11 [217]-[237].

52. Nine cases considered the foggy night analogy (beyond merely quoting the passage it appeared in): Debresay (n 49); Bayley v The Queen [2016] VSCA 160 (‘Bayley’); DPP v Wearn [2018] VSCA 39 (‘Wearn’); Bembo v R [2018] VSCA 42; Langford v Tasmania [2018] TASCCA 1 (‘Langford’); Cope v The Queen [2018] VSCA 261 (‘Cope’); Farhat (n 1); R v Salcedo (No 2) [2018] ACTSC 104; Dempsey (a Pseudonym) v The Queen [2019] VSCA 224.

53. Debresay (n 49).

54. Bayley (n 52); Wearn (n 52).

55. Bayley (n 52); For a review of Bayley, see Alena Skalon and Jennifer L Beaudry, ‘The science behind Bayley v The Queen (2016)’ (2019) 26(2) Psychiatry, Psychology and Law 219.

56. Bayley (n 52) [58].

57. Ibid [53].

58. See the sources at notes 38 and 39.

59. Wearn (n 52).

60. Ibid [47].

61. Ibid; the ACT Supreme Court also seemed to consider this factor, although in the context of unfair prejudice in R v Eastman (no 31) [2018] ACTSC 11 (31) [235].

62. Wearn (n 52) [47].

63. Ibid [25]-[26].

64. While the court in Wearn did couch contradictions and inconsistencies as reliability and credibility issues, a charitable view of this decision is that they are better characterized as going to credibility, and that is the reason they should not have been considered. Prior to IMM, even the more interventionist Victorian position was to consider credibility when assessing probative value: Dupas (n 17). As noted, however, there were factors separate from the inconsistent statements alleged in Wearn, like the possibility of publicity affecting the witness’s memory.

65. Debresay (n 49).

66. Ibid [60].

67. Ibid [15] (emphasis added).

68. Ibid [13], [40], [43], [46].

69. Ibid [47].

70. R v Afu [2018] NSWSC 190 (‘Afu’); Wearne may also be inconsistent with Cope (n 52) [42], in which the Victorian Court of Appeal seemed to accept the possibility that factors like the 30 year delay between the event and identification, and allegations of bias in the photoboard, could reduce the convincingness of the evidence (but ultimately admitted the evidence: ‘[e]ven if one assesses the probative value of the evidence as being low, the identification evidence taken at its highest has the limited capacity to rationally affect the assessment of the probability that the applicant was one of SD’s assailants’).

71. Afu (n 70) [17].

72. Ibid [20].

73. Ibid [21]-[22]. Note the published speech that the IMM majority was referring to did make reference the cross-racial identification effect, but this did not make it into the IMM judgment: Heydon (n 30) 234.

74. Farhat (n 1).

75. Ibid [39].

76. Ibid [39].

77. Ibid [41]

78. For context, 18 decisions considered unconvincing or simply unconvincing in the context of eyewitness evidence, whereas 28 decisions considered those phrases in the context of other types of evidence (eg expert evidence, tendency evidence).

79. Tasmania v Langford [2018] TASCCA 1.

80. Ibid [56].

81. For conditions under which an opinion might be so reliant on unproven assumptions as to be inadmissible, see Dasreef Pty Ltd v Hawchar (2011) 243 CLR 588.

82. Farhat (n 1) [38] [emphasis added].

83. Recall the factors in Debresay (n 49), such as the age of the witness and the opportunities for memories to become contaminated.

84. Known to whom? Here we can start to observe the problems created by a subjective standard — the persuasive to a judge — and more objective/independent concerns with reliability.

85. This also applies to fact-finders over-valuing their own comparison of the evidence. See Volpe v the Queen [2020] VSCA 268.

86. This is curious because similar factors were relied upon to erode the value of the identification in the foggy night example.

87. See Edmond et al, ‘Science and Justice. Thinking forensics: Cognitive science for forensic practitioners’ (2017) 57(2) Science & Justice 144.

88. A Daniel Yarmey, ‘The psychology of speaker identification and earwitness memory’ in RCL Lindsay, DF Ross, JD Read, and MP Toglia (eds), The handbook of eyewitness psychology, Vol 2 Memory for people (Lawrence Erlbaum Associates, 2017) 101.

89. Farhat (n 1) [40].

90. Ibid [40].

91. Curiously, they are not considered to introduce serious unfair prejudice to the defendant.

92. Importantly, these same factors may contaminate fact-finding at trial. If a recording of Farhat’s voice was available, the jury would have been provided with a copy to enable it to undertake its own comparison. Regardless, at trial jurors heard a voice recording of someone speaking with a Middle Eastern accent and were provided with a transcript that included the name ‘Shadi’ (the defendant’s name). They would have observed the defendant (perhaps of Middle Eastern appearance), sitting in the dock. They would hear the testimony of a senior police officer positively identifying one of the speakers as Shadi Farhat.

93. He also makes an unqualified identification. Not even DNA matches are expressed in such terms.

94. Yarmey (n 88). See also Smith v The Queen (2001) 206 CLR 650.

95. This is also a problem for the jury. In terms of trial procedure, it is not entirely clear how these issues could be effectively conveyed to the jury, especially where a police witness adheres steadfastly to his or her sensory impressions of the identity of a speaker.

96. Other cases have limited the possibility of considering reliability under other provisions, such as section 79 (the exception for expert opinion): R v Tang (2006) 65 NSWLR 681; Maxwell (n 7) 642-4.

97. Farhat (n 1) [38]; IMM (n 5) [50].

98. GS Morrison and WC Thompson, ‘Assessing the admissibility of a new generation of voice comparison testimony’ (2017) XVIII Columbia Science and Technology Review 326.

99. See R v Volpe (Ruling No 1) [2018] VSC 796 [84]; R v QL [2017] NSWSC 1253 [30].

100. This is a logical fallacy.

101. However, the converse may be true. Where an individual engaged in voice comparison, such as Sgt Bishop, is apprised of suggestive information that is not required for the analysis — such as the beliefs of other police — the resulting identification is contaminated — it is no longer independent evidence. See Edmond et al (n 87).

102. It is unclear how this should be applied. Should the probative value of the impugned evidence by considered in combination with all of the adverse evidence, or only in so far as the evidence informs the capacity of the impugned evidence?

103. Pfennig v The Queen (1995) 182 CLR 461 required the judge to consider the contested (similar fact) evidence in conjunction with all the evidence to determine admission. This contextual approach to tendency and coincidence evidence is also often followed under sections 97 and 98 of the UEL, see BC v The Queen [2019] NSWCCA 111 [75].

104. See the Victorian case of R v Crupi [2020] VSC 654. Crupi’s assessment of the reliability of gait evidence is an outlier: Emma Cunliffe and Gary Edmond, ‘Gaitkeeping in Canada: Mis-steps in assessing the reliability of expert testimony’ (2013) 92 The Canadian Bar Review 327.

105. Courts have refrained from reading reliability into the requirement that expert witnesses’ possess specialized knowledge under section 79 of the UEL, see Tuite v The Queen (2015) 49 VR 196 (‘Tuite’).

106. The Royal Society of London and The Royal Society of Edinburgh, Forensic gait analysis: a primer for courts (Royal Society, 2017) 6.

107. This would be like assuming that a new DNA-profiling technique, such as claims related to phenotype (eg determining eye colour or height from a DNA sample) work, without attending to the existence of formal support, that is, validation.

108. President’s Council of Advisors on Science and Technology, Forensic Science in Criminal Proceedings: Ensuring Scientific Validity of Feature Comparison Methods (Report to the President, September 2016) 6 (‘PCAST Report’).

109. Committee on Identifying the Needs of the Forensic Science Community, National Research Council et al, Strengthening Forensic Science in the United States: A Path Forward (National Academies Press, 2009) 188.

110. PCAST Report (n 108) 101.

111. Cunliffe & Edmond (n 104).

112. Compare forensic gait analysis (and other practices like latent fingerprint analysis and ballistics) to DNA analysis, which also produce types of identification evidence. We do not equate a matching DNA profile, even at its highest, with positive identification. Rather, the evidence is presented in a statistical form (eg the probability of a random match) that informs its probative value: Tuite (n 105). Bizarrely, categorical identifications are associated with latent fingerprints, some types of ballistics and toolmark evidence, and so on. But such categorical identifications have been consistently criticised by scientists. Scientists recommend that fingerprint examiners (and others) report the best available indication of performance: PCAST Report (n 108) 102.

113. Farhat (n 1) [41].

114. Ibid [38].

115. Ellis (n 14) [94]-[95].

116. We note that whereas probative value is required to be taken at its highest, unfair prejudice is assessed on the basis of the trial judge’s actual impression (however circumscribed); and even this is mitigated by the widespread judicial dogma that judicial directions can and will remove real dangers.

117. ‘I have already considered the issue of prejudice and unfair prejudice, and concluded that there is no real danger of unfair prejudice in the admission of the evidence, particularly bearing in mind the warnings required to be given to the jury.’: R v Fantakis [2018] NSWSC 1814 [80]; see also Gjonaj v The Queen [2018] NTCCA 13 [50]; Chen v The Queen [2018] NSWCCA 106 [75]; Jason M Chin, Mehera San Roque, and Rory McFadden, ‘The New Psychology of Expert Witness Procedure’ (2020) 42(1) Sydney Law Review 69, fn 21. We did identify two decisions in our sample in which the court did invoke ‘simply unconvincing’ to exclude the evidence, but seemed more persuaded by the prejudice of the evidence: R v Medich (No 8) [2016] NSWSC 1713 [59]; DPP v Lyons and Lyons (No 3) [2018] VSC 224 [74].

118. Cf Volpe v the Queen [2020] VSCA 268. Volpe is an impressive decision where the Victorian Court of Appeal applied s 137 to exclude shoe comparison evidence linking the defendant to a murder victim. However, the fact that the forensic analyst in that case proffered a modest opinion — a shoe found in the appellant’s premises ‘could have’ left a shoe print adjacent to the deceased’s body — meant that there was no speculative engagement with the value of the opinion at its highest. Section 137 was applied to fairly weak comparison evidence with the potential to mislead. Had the analyst offered a stronger conclusion (eg that the shoe made the print or that there is strong evidence that the shoe made the print), the Court may well have been much more constrained — in determining probative value, at least.

119. Dupas (n 17) [206].

120. ‘There is one particular factor which may affect the reliability of the identification made by Sgt Bishop which I think will, if the evidence is admitted, require a strong warning to the jury’: Farhat (n 1) [40].

121. Ibid [40], [44]. Revealingly, these directions are not informed by scientific knowledge about voice comparison or the most effective types of directions.

122. See Michael J Saks and Barbara A Spellman, The Psychological Foundations of Evidence Law (New York University Press, 2016) 87-90; Gary Edmond et al, ‘Forensic Science Evidence and the Limits of Cross-Examination’ (2019) 42(3) Melbourne University Law Review 858.

123. More generally, psychological studies regularly cast doubt on the effectiveness of judicial instructions, see: Roselle L Wissler and Michael J Saks, ‘On the Inefficacy of Limiting Instructions’ (1985) 9(1) Law and Human Behavior 37; J Alexander Tanford, ‘The Law and Psychology of Jury Instructions’ (1990) 69(1) Nebraska Law Review 71; Rachel K Kush and Jane Goodman Delahunty, ‘The Influence of Limiting Instructions on Processing and Judgments of Emotionally Evocative Evidence’ (2011) 13(1) Psychiatry, Psychology and Law 110, 112-13. Adding further complex instructions may only strain the cognitive resources of the jury: Supreme Court of Victoria, Simplification of Jury Direction Project Report (A Report to the Jury Directions Advisory Group), August 2012, 7-14.

124. What is also remarkable, reflecting on the case of Smith v The Queen (2001) 206 CLR 650, is that Bishop’s opinion evidence was construed as relevant. In Smith, a majority characterised police comparisons of images from the crime scene to their limited interactions with the defendant as irrelevant because the jury were in basically the same position as the police officers. In Farhat, we are dealing with even more error-prone voice identification where the identification was made in highly suggestive circumstances. Does the fact that the jury might not get to hear Farhat’s voice make Bishop’s subjective impression not merely relevant but highly probative? There are few reasons to think that a highly suggestive identification by a person engaged in the investigation is probative, let alone of high probative value. Unlike Smith, Farhat features the very serious complication of suggestion — an irreparable form of cognitive contamination.

125. It is not a credible response to rehearse the judicial belief that jurors understand and obey directions. The fact that these are not necessarily informed by scientific knowledge helps to convey the problem.

126. Bayley (n 52); Wearne (n 52); Debresay (n 49); Afu (n 70).

127. Maxwell (n 7).

128. See the sources at note 123.

129 Judges should be able to explain the capacity of the evidence in terms that go beyond the probative value being high and unfair prejudice low.

130. ‘On the present state of the law, therefore, the judge in a criminal trial is unable to perform the “gatekeeper” role as defined [U.S. and Australian decisions]. The purpose of this article is to underline the critical importance of that role, and to highlight the urgent need for legislative intervention to reinstate it.’: Maxwell (n 7) 643. This author was discussing reliability in the context of expert evidence, but as we have seen, more general reform would be needed to address reliability across the spectrum of evidence that s 137 once safeguarded criminal trials against.