Jurisdiction : SUPREME COURT OF WESTERN AUSTRALIA

Jurisdiction : DISTRICT COURT OF WESTERN AUSTRALIA

Appeal - Criminal law and procedure - Evidence - Warnings - Identification evidence - Digiboard/photoboard reliability - Relative identification - Witness' description differs from person identified - Effect of stress - Delay in identification

Criminal Appeals Act 2004 (WA), s 40(1)(e)

Evidence Act 1906 (WA), s 72

1 There is a significant overlap between these two appeals. In relation to Winmar, identification is the basis of the only ground of appeal, whereas in relation to Hussaini there is an identification ground and a ground based on an omission to give a "Liberato" direction (Liberato v The Queen [1985] HCA 66; (1985) 159 CLR 507). So far as the identification grounds are concerned, three of the particulars are identical. Those particulars assert that the dangers of identification in each case, against which it is submitted the jury should have been warned, included that there had been no identification parade and without such a parade the "subsequent photoboard identification was inherently dangerous"; and that photoboard identification was "undesirable and dangerous per se". The particulars in each case also assert that the jury should have been warned not to draw an adverse inference against the appellant simply because police had his photograph. In relation to each of the appellants there are also some further particulars, specific to each, raising other aspects of the identification warning, or lack of warning, with which we will deal later.

2 So far as photographic identification is concerned, somewhat different views have been expressed in this court about what a jury should be told concerning the allegedly unsatisfactory nature of photographic identification: compare Roser v The Queen [2001] WASCA 190; (2001) 24 WAR 254 at [64] per Anderson J with Collard v The Queen [2000] WASCA 417 at [77] - [78] per Miller J. Whatever the precise form of direction, it seems to have been accepted on all sides that photographic identification is necessarily significantly inferior to identification by means of an identification parade: Collard at [14] per Kennedy J and at [75] - [79] per Miller J. However, with some exceptions (eg Wallwork J in Collard at [62]), courts in this State have considered that photoboard identification should be admitted, but that there should be a warning of some kind about it.

3 Not only is the question of the desirability or otherwise of photographic identification (particularly when compared with an identification parade) an issue in this court, but material placed before us also demonstrates that views in other jurisdictions differ concerning the practice. The Commonwealth and New South Wales Evidence Acts provide by s 114 that visual identification adduced by the prosecutor is not admissible, unless an identification parade that included the defendant was held before the identification was made, subject to certain exceptions. Although Tasmania has adopted most of the uniform Evidence Act provisions, it has not adopted s 114. It appears from the materials placed before us that photographic identification is widely accepted in the United States; for example, the American Bar Association's Criminal Justice Section report to the House of Delegates (American Bar Association Statement of Best Practices for Promoting the Accuracy of Eyewitness Identification Procedures, August 2004) is concerned with setting out best practice guidelines for "a lineup or photospread" (ABA Statement, par A1) without suggesting that one or the other is a preferable means of identification. In the United Kingdom, we were informed, video identification procedures have been introduced.

Identification evidence: issues in this appeal

4 In considering what, if any, particular warnings should be given in relation to photographic identification, it is we think desirable to begin with first principles. We therefore first consider what authority requires in relation to warnings concerning identification evidence generally, and the reasons why such warnings are required. We consider what authority might suggest about the specific content of such a warning, and how, consistently with authority, this court can determine what the content of a warning in any particular case should be. We consider what role, if any, evidence concerning the dangers of particular types of identification might play in determining an appropriate warning.

5 Against that background, we then consider the evidence adduced by the parties in relation to the type of photographic identification which is routinely undertaken in Western Australia today, and consider what light the evidence adduced by the parties might shed upon the reliability of that form of identification. The result of that consideration is that, to the extent that there is authority binding on this court which deals specifically with the question of photographic identification, that authority can be seen to be concerned with photographic identification which was substantially and relevantly different from the "digiboard" identification in use in Western Australia, and is therefore distinguishable.

6 We then consider what warnings, if any, specific to digiboard identification should be given. We then turn to those aspects of each appeal which are not concerned with photographic identification.

7 The leading authority in Australia in relation to warnings concerning identification evidence remains Domican v The Queen [1992] HCA 13; (1991) 173 CLR 555. At 561 - 562, the majority said:

Whatever the defence and however the case is conducted, where evidence as to identification represents any significant part of the proof of guilt of an offence, the judge must warn the jury as to the dangers of convicting on such evidence where its reliability is disputed. The terms of the warning need not follow any particular formula. But it must be cogent and effective. It must be appropriate to the circumstances of the case. Consequently, the jury must be instructed 'as to the factors which may affect the consideration of [the identification] evidence in the circumstances of the particular case'. A warning in general terms is insufficient. The attention of the jury 'should be drawn to any weaknesses in the identification evidence'. Reference to counsel's arguments is insufficient. The jury must have the benefit of a direction which has the authority of the judge's office behind it. It follows that the trial judge should isolate and identify for the benefit of the jury any matter of significance which may reasonably be regarded as undermining the reliability of the identification evidence. (Footnotes omitted)

8 The court went on at 564 to state that mere repetition of counsel's arguments is an insufficient discharge of the trial judge's duty and that:

If the matters to which counsel has referred may reasonably be regarded as undermining the reliability of the identification evidence, the trial judge must direct the jury that they are bound to take those matters into consideration in determining whether they will rely on that evidence.

9 The passages quoted invite attention to two questions. The first is, what are the dangers of convicting on disputed identification evidence. The second is, how does a trial judge determine whether matters may reasonably be regarded as undermining the reliability of the evidence, so as to require the judge to lend the weight of his or her authority to a warning about those weaknesses.

10 What, then, are "the dangers of convicting on [identification] evidence"? The basal proposition is that there have been significant miscarriages of justice where an honest and confident identification witness has given evidence which is not accurate, and that the potential for such a miscarriage is a risk in most or many identification cases. Neither the witness' honesty nor the witness' confidence guarantees the reliability of the evidence.

11 That basal proposition derives (at least in part) from the experience of the courts. At [1345] of Cross on Evidence (7th Aust ed, 2004) there are mentioned a number of famous cases in which identification errors of an astonishing character have occurred. Further, any legal practitioner or judge who deals with civil or criminal cases will quickly discover that apparently honest eyewitnesses to an event can differ widely about important aspects of it. As the Devlin Committee Report points out at pars 1.24 and 4.25 (P Devlin, Report to the Secretary of State for the Home Department of the Departmental Committee on Evidence of Identification in Criminal Cases, HC 338 (London, HMSO, 1976)), cross-examination is a tool of limited usefulness in testing a witness' ability to recognise faces, and demeanour is not a useful guide to accuracy in such a case .

12 The danger, then, is that evidence which is inaccurate may be apparently convincing and that it will be difficult to test whether it is as accurate as it seems. The danger which must be warned against in every case, therefore, is the danger that an honest witness may be mistaken, and that an honest but mistaken witness may be convincing: Kelleher v The Queen [1974] HCA 48; (1974) 131 CLR 534 per Gibbs CJ at 550 - 551); R v Turnbull [1977] QB 224 (cited in Domican at 561); Festa v The Queen [2001] HCA 72; (2001) 208 CLR 593 at [166] per Kirby J, [218] per Hayne J; and Longman v The Queen [1989] HCA 60; (1989) 168 CLR 79 at 108 per McHugh J (dealing with honest but mistaken recollection generally).

13 There is another, related danger which authority suggests should be the subject of a general warning about the potential unreliability of identification evidence, at least in any case where there is any factual basis for concern that the danger may be present. It is the vulnerability of identification evidence to suggestion: Davies v The King [1937] HCA 27; (1937) 57 CLR 170 at 181 - 182; Festa at [22] and [26] per Gleeson CJ, at [78] and [81] per McHugh J. There are a number of aspects of this vulnerability to suggestion, including a possibility that a witness will pick out someone from photographs, or from a group of persons selected by the police, because he or she expects the group to contain the offender, and the possibility that a person may substitute in the person's memory an image of someone seen elsewhere (possibly in photographs selected or a parade organised by the police) for a hazy recollection of an offender.

14 Finally, it is obvious that a person's ability to observe an offender may be very limited in the circumstances of a particular offence, and the dangers of which a jury must be warned include, where appropriate, the danger that the witness will simply have not had an adequate opportunity to observe, so as to be able to identify anyone.

15 The matters outlined above seem to us to be the core warnings as to the dangers of convicting on disputed identification evidence which are recognised by authority. The first danger, which is that of accepting as accurate evidence which is convincing but inaccurate, is a danger present in every case. The others may well be present, to a greater or lesser degree, in most cases.

16 However, the dangers outlined above were obviously not the only matters which were of concern to the High Court in Domican. The "significant weaknesses" which their Honours discussed at 562 - 563 included, but were not confined to, difficulties of observation and possible suggestibility.

17 The precise circumstances of Domican are unlikely to recur. It is necessary therefore for trial judges to be able to identify those matters which may reasonably be regarded as undermining the reliability of identification evidence and about which a warning may be required. It is obviously impossible to lay down a rule applicable to every case, as circumstances are likely to vary.

18 However, it is, we think, desirable to give as much guidance as possible, when directions of trial judges are criticised for inadequacy in warning. A trial judge faces a difficult task in such a situation. He or she is required to give a jury any warning (as to identification or other issues) which may be called for by the particular case: Azzopardi v The Queen [2001] HCA 25; (2001) 205 CLR 50 at [49] per Gaudron, Gummow, Kirby and Hayne JJ. Further, it is recognised that it is not simply the function of the judge to repeat the arguments of counsel in relation to possible weaknesses in identification evidence. The weight of judicial authority must be placed behind the warning: Domican at 564 (and see Al-Hashimi v The Queen [2004] WASCA 61; (2004) 145 A Crim R 186 at [50] per Miller J).

19 However, it is of course the function of the jury to assess the facts. Unless a warning is required, then:

... it will often be better (and safer) for the judge to leave the assessment of the facts to the determination of the jury in the light of the submissions of the parties. Unnecessary or extensive comments on the facts carry well recognised risks of misstatements or other errors and of blurring the respective functions of the judge and the jury. (Azzopardi at [52])

20 This is an important caution. Judges do have particular expertise in relation to some aspects of the dangers of identification evidence. However, they are neither experts in the science of identification evidence, nor experts in the assessment of relevant literature in that field. They are not entrusted by law with the jury's function of evaluating the honesty and reliability of evidence in the light of ordinary experience. It would be wrong to assume that the wisdom of any individual judge (or even the collective wisdom of three or - as here - more appellate judges) is necessarily superior to the experience and understanding of 12 jurors.

How does a court determine a particular warning is required?

21 Consideration of the second question - that is, how a judge identifies the specific weaknesses about which it is necessary to warn - may be assisted by a consideration of the nature of judicial warnings generally. A review of the cases in Australia and England suggests that, before a warning is necessary, two factors must be present. First, there must be some aspect of the evidence which gives rise to "a perceptible risk of miscarriage of justice": Carr v The Queen [1988] HCA 47; (1988) 165 CLR 314 at 330 per Brennan J; Kelleher v The Queen at 560 per Mason J; Longman v The Queen at 86 per Brennan, Dawson and Toohey JJ.

22 Second, the risk of miscarriage of justice must be one which is not necessarily obvious to the lay mind (or perhaps one to which a lay person may give inadequate weight). It must derive from a factor "of which the judge has special knowledge, experience or awareness": Carr at 325 per Brennan J and at 341 per Gaudron J; Longman at 91 per Brennan, Dawson and Toohey JJ; Crampton v The Queen [2000] HCA 60; (2000) 206 CLR 161 at [132] per Kirby J; Chamberlain v The Queen (No 2) [1984] HCA 7; (1984) 153 CLR 521 at 604 per Brennan J.

23 If it is the special knowledge or awareness of the judge or of the courts which gives rise to the need for a warning, it follows that where a judge, or the courts, has or have no special expertise or knowledge, a warning is not only unnecessary, but is an unwarranted intrusion into the province of the jury: see Bromley at 324 per Brennan J; Carr at 321 per Wilson and Dawson JJ. In deciding what is or is not within the special knowledge or awareness of the court, there is a danger that the courts may assume an expertise which they do not possess. As the learned authors of Cross on Evidence (7th Aust ed, 2004) observe, although some perceived lessons from human experience are doubtless entirely correct, others may not be, and generalisations from experience may be modified as experience changes and may at any particular time be controversial (at par 3295). An obvious example is afforded by the rule which formerly applied, that in cases of an alleged sexual assault by a male upon a female, the judge was required to give a warning to the jury that it was dangerous to act on the uncorroborated evidence of the complainant. That rule was abrogated only by legislation, and not by a change to the common law, even though, as Deane J observed in Longman's case, it sounded "somewhat strangely in modern ears" (at 92) and even though, at least in his Honour's view, neither wisdom nor experience - judicial or otherwise - justified the rule at the time at which it came to be abrogated (at 93).

24 So far as intermediate courts of appeal are concerned, where a particular form of warning has been identified by the High Court as necessary in a particular case, there is no conceptual difficulty. The court is simply required to follow authority, and the only question which may arise is that of whether the circumstances of the case at hand do fall within, or are closely analogous to, the category of cases calling for the warning. The discussion which follows here is concerned with those issues concerning identification evidence which are not the subject of authority directly on point, but which may call for a warning.

25 The basic proposition which lies at the heart of the identification warning is a counter-intuitive one, which contradicts "common sense"; it is the proposition that a confident and honest (and, often, apparently careful) witness may, nevertheless, be entirely wrong. There is a risk that, in endeavouring to warn juries about other aspects of identification evidence, a judge may resort to what he or she knows or believes to be common sense, but that, in doing so, he or she may mislead a jury about what is (or is not) likely to indicate reliability or unreliability in identification evidence.

26 It may be appropriate to refer to scientific works, in some circumstances, in determining the content of a judicial warning, where the need for it is disputed. It appears that the identification warning itself is informed not only by the collective experience of the courts in decided cases; rather, it is likely that judicial awareness of other materials has helped to shape it. The courts of the United Kingdom have been significantly influenced by the Devlin Committee: R v Turnbull at 231 (applied in a long line of United Kingdom authorities; eg, Scott v The Queen [1989] AC 1242; Shand v The Queen [1996] 1 WLR 67). The Devlin Committee Report was referred to in Australia in Alexander (at 428, 430 per Mason J, at 436 per Murphy J). The Canadian courts have taken note of the Sophonow Report: P Cory, The Inquiry Regarding Thomas Sophonow (2001). For example, in Hibbert v The Queen [2002] 2 SCR 445, Arbour J made the following observation:

The danger of wrongful conviction arising from faulty but apparently persuasive eyewitness identification has been well documented. Most recently the Honourable Peter deC. Cory, acting as Commissioner in the Inquiry regarding Thomas Sophonow, made recommendations regarding the conduct of live and photo line-ups, and called for stronger warnings to the jury than were issued in the present case ...

While it is unnecessary to consider these recommendations in detail, I share the concern expressed by the Commissioner and, in this particular case, I think it would have been prudent to emphasize for the benefit of the jury the very weak link between the confidence level of a witness and the accuracy of that witness ... (469)

27 Works on evidence, consulted by legal practitioners and courts alike, discuss relevant research: eg, Cross at [1350], [1355]; G L Williams, The Proof of Guilt (3rd ed, 1963) at 109 - 110, 122. Scientific material has been referred to in order to give content to warnings in other circumstances: eg, Longman at 107 - 108 per McHugh J.

28 In the present case, a variety of material was placed before the court. There was affidavit material, prepared on behalf of the respondent, explaining what a "digiboard" is, how it is compiled, and how it is presented to a witness, together with material gathered from police forces within Australia, which showed the preferred procedure for identification in each jurisdiction. That material was admissible pursuant to s 40(1)(e) of the Criminal Appeals Act 2004 (WA). In addition, each party produced materials collected from books and journals concerned with scientific research into the process of identification. That material was admissible pursuant to s 40(1)(e), read together with s 72 of the Evidence Act 1906 (WA), which provides:

1. Authoritative texts on history, science etc., reference to

All courts and persons acting judicially may, in matters of public history, literature, science, or art, refer, for the purposes of evidence, to such published books, maps, or charts as such courts or persons consider to be of authority on the subjects to which they respectively relate.

29 There are obvious dangers in an appellate court referring to scientific material in this way. The works placed before it are of necessity only a selection of literature, often from a large field; the assertions in them have not been tested by cross-examination; and the works may often be of a technical nature and open to misinterpretation. There is also, of course, a danger that a court may take a particular view, based on evidence which is before it at a point in time, but that the material placed before it may be demonstrated by subsequent research to be inaccurate or unreliable. Where that occurs in relation to a court of appeal, there is a potential for many trials to be influenced by an erroneous view. However, where there is a clear consensus, which it appears unlikely that further study will alter in the foreseeable future, it seems to us that it is appropriate for the court to have regard to that consensus in determining whether a particular warning should or should not have been given. Even where there is not such a consensus, research which suggests that a particular line of reasoning may not be well-founded may lead the court to refrain from requiring a warning which it might otherwise have considered necessary.

30 At a directions hearing at which both of these appeals were called on for directions together, orders were made which facilitated the placing before the court, by the parties, of such material as they contended could shed light on the particular dangers alleged to exist in relation to identification evidence in each of these appeals. From the materials collated by the parties, there does appear to be a clear consensus of scientific view in relation to some aspects of identification evidence. All parties agreed that that was so. We turn to the substance of that consensus shortly.

31 We discuss here the compilation and physical appearance of the digiboard, and the process of showing it to witnesses as described in affidavits provided by the State. A digiboard is a series of photographs, usually 12 in number (the guidelines require a minimum of 12 photographs), in which a photograph of the suspect or accused person appears, together with a number of "fillers".

32 If photographic identification is considered possible, the investigating officer will advise the Forensic Imaging Unit that a digiboard is required, will nominate the suspect or person of interest and, if there is more than one photograph of that person in existence, nominate which is to be used in the digiboard. It appears that the Forensic Imaging Unit is a specialist unit, independent of the investigation. The Forensic Imaging Unit selects images of similar age, appearance, race, etc, to the suspect. Background colours and lighting conditions are matched. The photographs can be enhanced or altered in order to ensure the greatest possible similarity between the suspect and the fillers, for example, by lightening or darkening of hair or complexion.

33 It was not clear from the materials provided to us what course is followed where the suspect differs significantly in appearance from the description given by the witness; that is, whether the "fillers" were matched to the description given, or to the appearance of the suspect. The course taken in the case of Mr Hussaini suggests that a mixture of images is used.

34 The procedure for conducting a digiboard identification is that, once the board is prepared by the Forensic Imaging Unit, it is shown to the witness. Wherever possible, the person showing the digiboard to the witness will not be a person involved with the investigation. There is a standard set of remarks which the person showing the digiboard is required to make to the witness, and standard questions which are asked of the witness, all of which are on a form accompanying the digiboard.

35 The witness is advised that they will be handed a folder with numbered photographs and the witness is asked to look closely at all of them to determine if the witness is able to positively identify any of the persons shown. The witness is advised that there are a number of important points of which the witness needs to be aware. The form continues: "Firstly, an image of the person involved in the incident under investigation may or may not be in this folder. Do you understand me?" There is a space for the witness' answer to be recorded. The witness is advised that viewing the photographs contained in the folder does not mean that the witness is obliged to make a selection and, again, there is a space to record whether the witness understands that. The witness is advised that neither the person showing the photographs nor any other person is able to assist with the making of a selection, and that once the folder has been handed over, the person giving it to them will not speak to the witness again until such time as the witness has indicated one of the images, or has advised that they are unable to recognise any of the images. The witness is asked whether he or she has any questions, before the folder is handed over. There is a space on the form to record the words used by the witness when either identifying an image, or returning the folder without making an identification.

36 Once a witness has taken part in the identification procedure, the witness is asked not to speak about the procedure, as far as practicable, to any other witness, until all identification procedures have been conducted. The identification procedure involves only one prospective witness at a time. A fresh digiboard is used for each witness, with a different position assigned to the suspect's photograph, so that, even if there is conversation between witnesses, the possibility of contamination of the recollection of witnesses is reduced.

37 Wherever practicable, the entire identification procedure is recorded on videotape. Videorecording of the procedure took place in both matters the subject of these appeals, and we cannot recall an occasion in Western Australia in recent years when recording did not occur. It is the usual, but not invariable, practice for the prosecution to play the videotape of the identification at trial and to tender the videotape of the identification process as an exhibit.

38 At one level, the answer to this question is obvious; there is no authority to which we have been referred which deals directly with digiboards. However, a question does arise as to whether existing authority which deals with the question of photoboard identification has anything of relevance to say concerning the use of digiboards.

39 There is a line of authority which suggests that there are problems inherent in the photoboard procedure, and which suggests that photoboards are inherently inferior to identification parades. The line of authority in Australia derives from Alexander v The Queen [1981] HCA 17; (1981) 145 CLR 395. It is possible to identify from that case a number of concerns about photographic identification which were shared by all members of the court. There are also in that case some concerns mentioned by only some members of the court.

40 The absence of the accused from the identification process was seen in Alexander's case as a major problem with photographic identification: Gibbs CJ at 400, Stephen J at 409, Mason J (Aickin J agreeing) at 430, Murphy J at 436. The absence of the accused did not, of course, directly affect the reliability of the identification, but made it difficult for the accused to test the fairness of the identification process, and its likely reliability. The use of a standardised question format, which must be completed and retained, and the videotaping of the procedure, would appear to eliminate this concern in relation to the digiboard identification process.

41 The "rogues' gallery" effect was of concern in Alexander to Gibbs CJ (at 400 - 401), Stephen J (at 409), Mason J (at 426) (Aickin J agreeing) and Murphy J (at 436). The "rogues' gallery" concept needs to be unpacked a little. It was certainly the case in earlier times, in the experience of this court, that photographs shown to witnesses looked very like the lay idea of a "mug shot" (perhaps because that was what they invariably were). They were generally photographs of sullen-looking individuals in what appeared to be institutional clothing, in what appeared to be an institutional setting, under harsh lighting. At the time at which Alexander was decided, photographs of that kind would sometimes, in this State, even display or allow a glimpse of numbers, which one might well conclude were numbers assigned to prisoners. Digiboard images, by contrast, have a neutral background and do not display institutional clothing, or numbers.

42 Another aspect of the "rogues' gallery" effect, is the not unreasonable assumption that a jury may consider that the police have taken or obtained the suspect's photograph because he or she is suspected of committing the offence in question. In the ordinary case, this consideration is not likely to be of any importance; it is obvious, from the fact that the accused is on trial, that he or she is suspected by someone of having committed the offence.

43 The real difficulty with the "rogues' gallery" effect arises where an accused person is identified from photographs prior to being questioned or arrested. The assumption may readily arise that the photograph was taken on the occasion of an earlier arrest or conviction. However, it is in precisely these circumstances that an identification parade cannot be arranged, so that the prejudice is to that extent unavoidable. There has been discussion in the cases, of course, concerning the desirability of a warning in these circumstances about drawing any inference from the fact that police apparently have a photograph of the accused.

44 The "displacement effect" was referred to in Alexander (at 409, 426, 436). However, the "displacement effect" discussed in that case appears to have been referring to the phenomenon which could occur when a witness, who had first picked an accused person from photographs, later purported to identify the person at an identification parade. The witness might be identifying the accused at the parade, not as the person seen at the time of the offence, but as a person previously identified in a photograph, and a jury might give undue weight to the subsequent identification. Where there is no subsequent identification parade, that effect does not occur. There is also, of course, the possible "displacement effect" of the type discussed in Davies and Cody, which can occur when a witness is shown only a photograph of the suspect, or only the suspect in person; that issue is not relevant here. There are other circumstances in which there can be a "displacement effect", one of them relevant to the appeal of Winmar, but none of them is more likely to occur in relation to photographic or digiboard identification than in relation to other forms of identification.

45 The discussion above exhausts, we think, the list of factors, each of which was identified as a matter of concern by several members of the court in Alexander. However, Stephen J, in that case, identified other potential disadvantages of photographs which have been echoed in later cases. His Honour referred, at 409, to certain peculiar difficulties relating to photographs, being "their two dimensional and static quality, the fact that they are often in black and white and the clear and well lit picture of the subject which they usually provide". Some of those concerns have been removed or alleviated by technological change. Digiboards are not black and white and, as we have noted, it is possible to adjust the way in which they are lit.

46 It may be that, in the future, technological change will remove or alleviate the difficulty which arises from the "static" nature of a photograph. We were informed by the State that a form of video identification is in use in the United Kingdom, and some of the materials before us referred to experiments in which a person who is asked to identify another is able to rotate the image or images being scrutinised, so as to view the representations from a variety of angles: see J Bailenson et al, "The Effects of Witness Viewpoint Distance, Angle, and Choice on Eyewitness Accuracy in Police Lineups Conducted in Immersive Virtual Environments" (2006) Presence 72. However, that is not the present state of the technology in Western Australia.

47 However, returning to the current digiboard process, it is true that it is a two-dimensional image; it is also the case that the image shows only the head and not the whole body, so that potential identifying or exclusionary factors such as height, build, and posture are removed. To that extent the digiboard process has its disadvantages.

48 Where a digiboard is used, it would therefore be desirable, in our view, that the judge generally warn the jury that they should take into account the fact that the digiboard is a static and two-dimensional image. Whether it is necessary to mention other possible disadvantages will depend upon the circumstances of the case. For example, in the trial of Hussaini, the evidence was that the witness had only seen the offender when the offender was sitting in a van. She was able to give evidence about the colour of his shirt, but it appears that she had no opportunity to observe his whole body, so that there is no question of matters such as build and posture entering into the identification process.

49 Turning, then, to the broad question of whether, as each appellant submits, a jury should be warned that digiboard identification is "undesirable and dangerous per se", the only significant disadvantages of identification by digiboard relate to the two-dimensional and static nature of a photograph, and to the fact that it is a head, or sometimes head and shoulders, representation only. The significance of these drawbacks will vary from case to case. There are, however, a number of advantages of the digiboard identification, as compared to an identification parade.

50 It may be that the more closely the "fillers" resemble the suspect, the more satisfactory the identification process (Alexander per Mason J, 428, quoting from the Devlin Report). This proposition is not universally accepted; there are studies which suggest that selecting fillers who resemble the suspect can have a negative impact on the reliability of the identification. (See G Wells et al, "Eyewitness Evidence: Improving Its Probative Value" (2006) 7(2) Psychological Science in the Public Interest 45, 62; D McQuiston-Surrett et al, "Sequential vs Simultaneous Lineups: A Review of Methods, Data, and Theory" (2006) Psychology, Public Policy and Law 1, 8.) However, most of the materials before us assume that matching either the suspect, or the description, is desirable. If that is so, then digiboards may assist in a reliable identification. Because of the ability to manipulate the images in the digiboard, a closer resemblance to the accused may often be achieved by that process, than by an identification parade.

51 It is plainly undesirable that anything in the circumstances of the identification process suggests to the witness that one person is more likely to be the offender (Alexander per Gibbs CJ, 400). Some of the material which has been placed before us suggests that one potential difficulty with an identification parade is that the suspect's demeanour may be different from the demeanour of the fillers, either because the suspect knows him or herself to be guilty, or understands him or herself to be the suspect, or perhaps because the fillers, if all or some are regular volunteers, are more likely to be relaxed: see J Turtle et al, "Best Practice Recommendations for Eyewitness Evidence Procedures: New Ideas for the Oldest Way to Solve a Case" (2003) 1 The Canadian Journal of Police and Security Services 5, Note 3; Y Tinsley, "Even Better Than the Real Thing? The Case for Reform of Identification Procedures" (2001) 5(99) International Journal of Evidence and Proof 1, 2, 4. Further, if stress can affect identification, it may be less stressful (especially if the offence has been a violent one) for the witness to view photographs only.

52 It seems to be generally accepted that some minimum number of fillers is necessary for a satisfactory identification process: see G Wells et al, "Eyewitness Evidence: Improving Its Probative Value" (2006) 7(2) Psychological Science in the Public Interest 45, 62 - 63. In practical terms, the digiboard offers the opportunity of obtaining a larger number of representations of people resembling the suspect than would be obtained by asking individuals to attend for an identification parade.

53 Finally, it is accepted that memory generally deteriorates over time: see G Wells et al, "Eyewitness Evidence Improving Its Probative Value" (2006) 7(2) Psychological Science in the Public Interest 45, 54; American Bar Association Statement of Best Practices for Promoting the Accuracy of Eyewitness Identification Procedures (2004) 9; N Brewer et al "Eyewitness Identification", in N Brewer and F Williams (eds), Psychology and Law: An Empirical Perspective (2005) 177, 191 - 192. Because the digiboard is relatively quick to organise, it reduces (compared to an identification parade), to a greater or lesser degree, the effect of time upon memory.

54 To the extent that research deals directly with the question of whether the digiboard may be inherently inferior as a means of identification, a review of a number of studies analysed in 1994 concluded:

... based on available research, there is no reason to believe that live lineups, videotaped lineups, or photo arrays produce substantial differences in identification performance. Based on what is currently known, identifications from photo arrays should therefore not be given less weight in investigations or in trials than identification from live lineups.

(B Cutler et al "Conceptual, Practical and Empirical Issues Associated with Eyewitness Identification Test Media" in D Ross et al (eds) Adult Eyewitness Testimony (1994) 163, 181.)

55 Having regard to what is currently known about the difficulties affecting identification, and having regard to the nature of the digiboard process as we have described it, it is our view that this court should firmly reject any suggestion that the digiboard process is inherently inferior to an identification parade. The court should not, as some past authority may tend to suggest, attempt to discourage the use of the digiboard for identification, either by requiring trial judges to warn juries specifically about the dangers of that process as compared to an identification parade, or by requiring trial judges to suggest to juries that the process is inherently flawed, or by suggesting that trial judges should be readier, in the exercise of their discretion, to exclude digiboard identification than they might be to exclude evidence of identification by other means.

56 Obviously, we do not intend to suggest that trial judges should suggest to juries that there is anything particularly reliable about the digiboard process. All identification evidence carries with it the particular danger we have earlier discussed; that is, the danger that, where the offender was not previously known to the witness, the witness may honestly, but mistakenly believe that they have identified the correct person, with the risk that the jury will accept such mistaken evidence as accurate. However, that danger is not invariably increased, and may in some cases be lessened (as compared to other methods of identification) by the digiboard procedure.

Relative judgment and sequential identification

57 Before we turn to the features peculiar to each of the individual appellants, it is desirable to deal with one issue which the research material has thrown up, and which is common to both cases. Although the question was raised directly only in Hussaini's case, submissions were made by counsel for both appellants concerning the phenomenon known as "relative judgment". It appears to be generally accepted that there is a risk that eyewitnesses who view all of the participants in an identification parade, or the photographs on a photoboard or digiboard, simultaneously, may make a positive identification by deciding which of the participants presented to them most resembles the offender. The problem is that false identifications may be made because there will always be a person in the digiboard or identification parade who looks more like the offender than the others, even when the offender is not present: G Wells et al, "Eyewitness Identification Procedures: Recommendations for Lineups and Photospreads" (1998) 22(6) Law and Human Behaviour 1, 10; G Wells "Eyewitness Identification: Systemic Reforms" (2006) Wisconsin Law Review 615, 619). There is some support for the view that a preferable procedure is that of a "sequential line-up" in which the witness is shown one person, or one image, at a time, in order to encourage the witness to make absolute rather than relative judgments.

58 However, there also appear to be some drawbacks of the sequential method. It appears that sequential methods may lead a witness not to identify the offender when the offender is present, an inaccuracy favourable to an offender, but unfavourable to the community: American Bar Association, Criminal Justice Section, Report to the House of Delegates: Recommendation IIIC (2004) 11.

59 It also appears that some studies suggesting increased accuracy for the sequential method may be reflecting, at least in part, increased accuracy in the "double-blind" identification (that is, where the person showing the witness the images does not know who the suspect is): American Bar Association at 11. The "double-blind" is, wherever practicable, the method presently used with the digiboard. Additionally, a warning that the offender might not be present appears to reduce very significantly the relative identification phenomenon, without apparently significantly reducing accurate identifications when the offender is present: G Wells et al, "Eyewitness Identification Procedures: Recommendations for Lineups and Photospreads" (1998) 22(6) Law and Human Behaviour 1, 11.

60 It is also to be noted that a number of the studies are American studies and it appears that "line-ups" in the United States typically involve five or six participants: American Bar Association, 12. The size of the "lineups" in the studies is not clear. Since there appears to be an association between size and accuracy, in the sense that a smaller number of fillers is more likely to result in a misidentification, there is a question of the extent to which the relative identification phenomenon is likely to be reduced in a procedure involving a larger number of persons.

61 Finally, there is also, at least amongst some researchers, a view that the research is not yet sufficiently mature to be able to determine whether, or in what circumstances, the sequential procedure is to be preferred to the simultaneous procedure: American Bar Association, 11. Indeed, it appears that sequential line-ups can be an inferior mode of identification in some cases. For example, it appears that young children may be more accurate in identification if they are first encouraged to make a relative identification, and are then asked whether the person they have selected as most resembling the offender is, in fact, the offender: A Freire et al "Lineup Identification by Children: Effects of Clothing Bias" (2004) 28(3) Law and Human Behaviour 339, 353.

62 Having regard to the considerations we have outlined, it seems to us that the question of whether an identification is performed by showing images or persons sequentially, or all at once, is not of such critical importance as to require a warning, as a rule: see G Wells et al, "Eyewitness Identification Procedures: Recommendations for Lineups and Photospreads" (1998) 22(6) Law and Human Behaviour 1, 31. However, it may be that in some cases something said or done by the witness gives rise to a concern about whether the witness has performed a relative identification, rather than actually identifying the offender. Where that is so, it would be appropriate for there to be a judicial warning that the jury should take into account the risk that such an identification may have occurred. It would not, of course, be necessary to couch it in scientific, or any particular, terms; the risk that a person could pick the person who "looks most like" the offender, rather than the actual offender, is capable of being grasped if explained quite simply. There is nothing in either of the appeals under consideration here to give rise to a concern of that kind, however.

The relevance of description of the offender

63 One issue which was common, in a sense, to both appeals was the question of the relevance of "description": what, if anything, should be said to a jury about the absence of any record of a description of the offender prior to the identification process (if there was no such record); and what, if anything, is the relevance of any discrepancies between the description given and the person identified. So far as Winmar was concerned, it was pointed out during the course of submissions that there was no evidence that there was a description or identikit photograph, or anything of that kind given or made by the witnesses prior to the identification of the appellant. So far as Hussaini was concerned, there had been a description and an identikit image prepared, and there were significant differences between the description/identikit and the photograph which the witness identified as the offender. In particular, the image identified as the offender had a goatee beard, whereas the identikit did not. The hair style was a little different and the style of moustache was arguably different. Although his Honour the trial judge in Hussaini's case had, in the context of warning the jury about the need to scrutinise the identification evidence, told them that a relevant question was what description had been given of the offender, the absence of the goatee from the description was pointed out only during that portion of his Honour's direction which dealt with the addresses of counsel and reminded the jury of them.

64 It was submitted that it was important for the jury to be warned that they must consider whether there had been an early description of the offender by the witness, or whether there was a discrepancy between the description and the appearance of the person identified. The research material tends not to support the proposition that this is a factor which is significant in assessing the accuracy of identification evidence. The New Zealand Law Commission Evidence Miscellaneous Paper 13 (1999) summarised (at pars 66 to 70) a variety of studies, the overall result of which seems to be that there is, at most, a very weak relationship between the quality of the description provided and the accuracy of the witness' identification. Some studies appear to find a relationship between errors in description and identification performance, while others do not: N Brewer et al, "Eyewitness Identification" in N Brewer and F Williams (eds), Psychology and Law: An Empirical Perspective (2005) 177, 193.

65 As we understand it, there are a number of factors which might explain why the counter-intuitive suggestion of a lack of correlation between accurate description and accurate identification may be well-founded. Generally, it is considered that most people do not have good memories for isolated facial features, so that a description which is built up by describing particular features (even if done by way of an identikit process) may result in an overall picture which is not closely related to the person described: G Wells et al, "Eyewitness Evidence Improving its Probative Value" (2006) 7(2) Psychological Science in the Public Interest 45, 64 - 65. A witness in Hussaini's case gave a striking example of this process. When cross-examined about whether Hussaini had ever had just a moustache (rather than moustache and goatee), she said: "I had a husband with a mo. When he shaved it off I didn't even know it was gone" [green - AB 65].

66 The process of describing something involves trying to recall it and to put what is recalled into words, while the process of recognition is a different and, it is considered, relatively "holistic" process. This may be another factor accounting for differences between recognition and description: K Lyle and M Johnson, "Effects of Verbalisation on Lineup Face Recognition in an Interpolated Inspection Paradigm" (2004) 18 Applied Cognitive Psychology 393, 401. Further, a number of studies suggest that there appears to be a relationship between speed of identification and accuracy, so that a witness who quickly picks out an image of a person (often saying that the face "just jumped out" at them, or words to that effect) will more often be correct than a witness who takes a longer time. It is postulated that this is because recognition tends to be a largely automatic and unconscious process; witnesses who identify in this way often cannot really say how they chose the person, or what features they relied upon: N Brewer et al, "Eyewitness Identification" in N Brewer and F Williams (eds), Psychology and Law: An Empirical Perspective (2005) 177, 204 - 205; G Wells et al "Eyewitness Evidence Improving its Probative Value" (2006) 7(2) Psychological Science in the Public Interest 45, 68.

67 These studies, then, tend to support the proposition that description and recognition are cognitively different functions and that an identification may be accurate even if it does not match a description. Indeed, they tend to suggest that, even where, as is often the case, a witness appears unable to explain why it is that they chose a person who does not match their description in some respect, their inability to do so may have no bearing upon whether or not their identification was accurate.

68 Finally, of course, the common experience of the courts is that, in relation to matters such as distance, height, weight, and even colour, many people are not good at providing accurate estimates or clear descriptions, even where there is no reason to doubt their ability to recall what they are purporting to describe. To that extent, the experience of the courts supports the proposition that such discrepancies are not, or are not always, of significance. Of course, as a matter of obvious common sense, some discrepancies must be of significance; a witness who gives an age estimate of 25 and picks out a suspect of 60 must be doubted for that reason alone. But discrepancies of this kind are relatively rare, and are not in issue in either case before us.

69 It is not the role of a trial judge to explain to a jury that discrepancies between description and the appearance of the person identified, upon which defence counsel may have focussed, may not necessarily be a particularly useful guide to the witness' reliability, even if current research suggests that such discrepancies may be unimportant. It is not the function of a trial judge to instruct a jury how they may reason towards guilt, rather than to warn them against impermissible modes of reasoning: RPS v The Queen [2000] HCA 3; (2000) 199 CLR 620, [43] (Gaudron ACJ, Gummow, Kirby and Hayne JJ).

70 However, we do not think it is appropriate for this court to suggest that a judge is in error for failing to lend the weight of judicial authority to a proposition which, while it may appear to be "common sense", may lead a jury erroneously to focus on a factor of limited, if any, importance. It is therefore our view that neither the absence of evidence concerning any prior description (in relation to Winmar) nor the discrepancies between the description given by the witness and the appearance of the person identified (in relation to Hussaini) should have been the subject of any judicial warning.

71 So far as Hussaini is concerned, there is a further reason for taking the view that his Honour did not err in failing to deal with the question of the possible discrepancy between the description given, and the image identified. Even if, contrary to the view we have expressed, such discrepancies should have been the subject of a warning, the complainant in Hussaini's case had participated in the making of an identikit image, based upon her description. As we later describe, a neighbour of Hussaini's contacted police because of what he called the "striking similarity" between the identikit and Hussaini. We have not seen Hussaini in person, but the jury did. Any discrepancy between the photo and the description would not have assisted him, if the identikit in fact bore a close resemblance to him, and a warning may well have served only to emphasise the resemblance. That may well be why trial counsel gave limited attention to the issue of discrepancy, and did not seek such a warning. Finally, it should be noted that, although he did not "warn" about the discrepancy between the description and the photograph, his Honour did, in the context of indicating that he would mention aspects of the complainant's evidence which might affect the jury's assessment of her identification, draw the jury's attention to the description which she first gave to police (green appeal book 165).

72 The short answer to the proposition that a warning should have been given by the trial judge concerning the drawing of any adverse inference against the accused because the police had his photograph, in either of these cases, is that the evidence was in each case that a photograph was taken of each appellant at the time of arrest, and that the identification took place on either the same or a subsequent day. So far as Hussaini was concerned, the investigating officer gave evidence that the digiboard used a photograph he had taken of the appellant immediately after interview on 24 November 2004 (blue/green AB 47 - 49). So far as Winmar was concerned, the investigating officer gave evidence that he took a photograph of the appellant on 5 December 2003. The digiboard was shown to the complainants on 23 December 2003 (blue/green AB 102 - 107, 115). Any warning to the jury in either case not to draw an adverse inference, or not to speculate, about matters which were in each case the subject of evidence would have been not only superfluous, but likely to confuse.

73 Authority supports the view that a "rogues' gallery" warning is not required in cases such as these. In Festa, the "rogues' gallery effect" was not considered to be relevant because there was evidence before the jury that the photoboard had been prepared after the appellant had been charged (at [9], [22] per Gleeson CJ). Similarly, in R v Campbell [2007] VSCA 189, the court held that it was not necessary for the trial judge to provide a "rogues' gallery" direction because the evidence was that the photograph of the accused was taken on the same day on which he was arrested and interviewed by police and there was no suggestion that the accused had been previously known to police ([44] - [45]). Finally, in DPP (Tas) v Lynch [2006] TASSC 89; (2006) 166 A Crim R 327, it was held that the fact that the police had a photograph of the suspect could not have suggested to the jury that he had a criminal history because there was available evidence that the photograph was taken weeks after the alleged crimes ([27]).

74 We now turn to those aspects of Winmar's appeal which are peculiar to him. In order to understand them, it is desirable to describe, briefly, the evidence given at his trial.

75 Winmar was tried on an indictment containing one count of aggravated armed robbery and one count of aggravated burglary, both alleged to have occurred on 5 December 2003. The complainants in both counts were Mr Brankovic and Ms Connett. At just after 7.10 am on 5 December 2003, Brankovic had just finished a night-shift at work and was at home in bed, while Connett, who was at the time his fiancee, was getting ready for work and had just returned from the shower to the bedroom. They were at the time sharing the house with a Mr O'Donnell, who was not present. Connett heard a noise and looked down the passage, thinking that it was O'Donnell, but saw the back of an Aboriginal person. She retreated into the bedroom. The offender, who was not wearing a shirt, entered the bedroom. His right hand was covered with a cream-coloured shirt and he told them he was armed with a gun and demanded property.

76 He took some property in the bedroom, told the complainants to follow him to the lounge room and to other areas in the house, where he picked up, took and moved various items. After a little time, he told them to stay in the bathroom. Connett's mobile phone was in the bathroom and she dialled '000'.

77 Both Brankovic and Connett noticed tattoos on the offender's chest. The appellant had tattoos in similar areas.

78 The offender had told the complainants to count to 80 before coming out of the bathroom and that if they came out before that, he would shoot them. After they had waited a short while, Brankovic went out looking for the offender and from the front of his house noted a car accelerating away from a driveway across the road. He made a note of the number and approximately three and a half hours later, that vehicle was stopped by police.

79 In the vehicle which was stopped, there were three females and the appellant. In the interior and in the boot was property belonging to Connett. A cream-coloured shirt was recovered from the car. On 23 December, 18 days after the offence, Brankovic picked the appellant out as the offender from a digiboard. Connett was shown a different digiboard on the same date and picked out a person who was not the appellant, and the evidence was that there was no possibility that that person could have been at the complainants' house on the date of the offence.

80 The appellant's evidence at trial was that on the day in question he had been staying at his cousin's, which was near the complainants' property. In the morning, he saw two boys trying to break into the car he was using. He chased them and they dropped a bag, which he picked up. He saw Brankovic out the front of his property and waved at him.

81 His Honour told the jury that unless they were satisfied beyond reasonable doubt as to the reliability of Brankovic's identification of the appellant as the offender, the remaining evidence on its own would be insufficient to convict the appellant. He went on to say that in those circumstances, the law required him to warn the jury to assess that evidence of identification "very carefully and to be very cautious before accepting it as being reliable and accurate" (blue/green AB 184). Having said that there was a good reason for the warning, his Honour went on to explain that judicial experience over many years had shown that there could be mistakes in identification. His Honour said that sometimes honest and convincing witnesses will identify an accused as an offender, but be mistaken, and that that was particularly so if the witness had little opportunity to observe the alleged offender. He told the jury that they would need to assess very carefully all the circumstances in which the identification was made and to have regard to all of the factors that might bear upon the reliability of Brankovic's identification.

82 Having given that warning, his Honour identified certain relevant factors. They were, the period of time over which Brankovic was able to observe the offender; the lighting; the distance between Brankovic and the offender; and what distinctive features there were about the offender that Brankovic was able to observe. In respect of each of those factors, his Honour referred the jury to relevant evidence dealing with that issue. He advised the jury that they should have regard to how much time had elapsed between the observation and the identification.

83 His Honour made what was, in our view, an unnecessarily detailed set of observations concerning photoboards and the controversy surrounding their use. He noted that the defence took issue with the colouration of the photographs and that that was a matter for the jury to consider.

84 His Honour told the jury that they should have regard to what was called the "displacement effect" and told them that it was a recognised phenomenon that the brain could sometimes subconsciously substitute the memory of another person for the memory of the offender. His Honour explained how that might occur, if Brankovic had seen the appellant in the street very shortly after the offence had been committed.

85 It is clear, from that summary, that his Honour gave a detailed warning, which was appropriate to the circumstances. Apart from our reservation in relation to the photoboard issue (as to which his Honour's observations were understandable in the light of earlier cases in this court), we would respectfully describe it as a good example of a clear and detailed warning. There are, however, three aspects of it which are the subject of criticism.

86 First, as to what his Honour did not say, it is submitted that he should have told the jury to consider the possible effects of the stress of the events in question upon the complainants' ability to observe and to remember the offender. A submission to that effect, which defence counsel had made to the jury, was referred to by his Honour, but it is submitted that he should have put the weight of judicial authority behind that point.

87 In our view, the effect of stress upon a person's ability to identify another is not so clear that a judge should be required to direct upon it. In experimental circumstances, stress tends to reduce correct identification, but there is some evidence that in real crimes, the person who is under the most stress is usually the most accurate (perhaps because they are closer to the crime and have a better view): G Wells et al, "Eyewitness Evidence Improving its Probative Value" (2006) 7(2) Psychological Science in the Public Interest 45, 49. Further, the effects of stress are complex. On the one hand, it appears that a degree of stress may assist in focussing attention, and stressful events are more likely to be frequently remembered, thus strengthening the memory. On the other hand, extreme levels of stress may impair memory, or may cause a witness to focus on a few details only: see generally, M Kebbell and G Wagstaff "Face Value? Evaluating the Accuracy of Eyewitness Information" (1999) Police Research Series, Paper 102 (Home Office Policing and Reducing Crime Unit, UK) 11 - 12. While, in a particular case, the evidence of a witness about the effect of stress upon them may suggest that a judge should direct a jury to take the possible effects of stress into account, it does not appear to us that this is a direction which should be given as a matter of routine. There is nothing, in our view, to suggest that it was required in the present case.

88 During the course of describing relevant factors to which the jury should have regard, his Honour said: "It has been suggested you should also have regard to Mr Brankovic's state of alertness at the time of this incident ... " (blue/green AB 437). The submission is made that his Honour should have directed the jury to take into account the fact that Brankovic was "drowsy". That, however, was not Brankovic's evidence. He said that he was a bit drowsy at the time that his fiancee told him that she had heard a noise (blue/green AB 21, 46). However, as to his state of mind during the events in question, he said: "I was actually quite alert. If someone comes into your house and says those sort of things to you, you'd be quite alert too" (blue/green AB 64). If his Honour had given the direction contended for by the appellant, he would have been in error.

89 Finally, it is submitted that the trial judge should have told the jury that Connett's misidentification may have reasonably cast doubt upon the positive identification by Brankovic, since both complainants had equal time and opportunity to observe the offender and also, it is submitted, that Connett was "fresher", having just got up to get ready for work. What his Honour said about this matter was as follows:

What is important about that is that she did not identify the accused as the offender even though she had had the same opportunity as Mr Brankovic to observe the intruder. The defence submits that this supports the contention that Mr Brankovic's identification was incorrect or unreliable. It is a matter for you but I suggest that the degree of similarity or the degree of difference in appearance between the accused's photograph and the person selected by Mrs Brankovic [Connett] is relevant.

You will remember the evidence is, I think, that the accused was photograph number 7 on her photoboard and she selected photograph number 6. Now, I suggest you should look at those two photographs to see how similar they are and I suggest the degree of similarity or the degree of difference is relevant to the question of whether or not Mr Brankovic's identification is reliable.

For example, and to take an extreme example, if photograph 6 was a photograph of the accused's twin then you would say, 'Well, there's the explanation and it's very easy to understand the mistaken identification.' Similarly, if that mistakenly identified person was very similar in appearance to the accused you might well think, 'Well, that would explain that mis-identification and therefore it doesn't detract much from Mr Brankovic's identification.'

If on the other hand the photograph number 6 was very different to the accused in appearance then that may well lead you to the conclusion that you shouldn't place great reliance on Mr Brankovic's identification. That's just a matter of commonsense but that's, as I say, a matter for you. (blue/green AB 188 - 189)

90 It is also submitted, as to these paragraphs, that his Honour should have told the jury that the photographs identified by Connett and Brankovic were dissimilar to each other, so that doubt was cast upon Brankovic's identification.

91 In one sense, his Honour's direction was unduly favourable to the appellant, in suggesting that the failure by Connett to identify the appellant was a factor that might cast doubt upon the reliability of Brankovic's identification. In one sense, it might be said that logically each identification was an expression of opinion based upon the recollections and perceptions of the individual who purported to make the identification, and that a mistake by one would say nothing about whether the other was reliable. Rather, a mistake by Connett might raise a doubt about whether, on the prosecution case taken as a whole, it had been established that the appellant was the offender.

92 However, it appears to us that the jury would well have understood the point which his Honour was endeavouring to make. That was the simple point that if two people had had the same opportunity to identify the offender, and one of them had been mistaken in doing so, then that vividly illustrated difficulties attendant upon the identification process, and would lead the jury to consider even more carefully whether either of them really had the opportunity, or the ability, to make a reliable identification. The additional comments he made relating to the similarity or dissimilarity in photographs were a development of that proposition. If two individuals had the same opportunity to observe an offender, and the two observers then each identified individuals who looked strikingly different from each other, then, in the absence of special circumstances peculiar to one of the witnesses, it would be easy to conclude that the whole identification process was hopelessly flawed. If the two individuals identified were very similar to each other, then, although the misidentification would still illustrate the dangers of the identification process, one might less readily conclude that misidentification by one of them meant that the other one could not have had an opportunity to observe or to remember.

93 Although the appellant referred to the decision of this court in Mule v The Queen [2002] WASCA 101, that case concerned very different circumstances. In Mule, one witness gave evidence that she was confident, having looked at a photoboard which did contain a photograph of the accused, that the offender was "not there". That is, her evidence was, in effect, that the appellant was not the offender. The court in Mule considered that the jury should have been directed that they could not convict the accused in that case if they accepted the evidence of that witness, or if they could not exclude it beyond reasonable doubt. Connett's evidence was not that the appellant was not the offender, although it was that she thought the offender was a different person. The uncontested evidence was that Connett must have been in error. It followed that her evidence must have been excluded beyond reasonable doubt, so that a Mule direction would not have been necessary.

94 The above completes the discussion of the ground of appeal in Winmar. Winmar has not yet been granted leave to appeal on any ground. We would grant leave in relation to the ground as set out in the appellant's substituted case dated 14 May 2007, but would dismiss the appeal.

95 In relation to Hussaini, there are two further identification issues. First, it is submitted that his Honour should have specifically warned the jury about the delay between the offence and the date of identification, and, secondly, it is submitted that the identification procedure adopted in Hussaini's case was inherently dangerous, as the digiboard was constructed in a manner which made the appellant's photograph stand out. Although it is phrased in a way that suggests that it goes to the admissibility of the digiboard, as we understand the latter submission, it is suggested that there should only have been a "strong warning" about that deficiency. To put those submissions in context, we turn briefly to the case against Hussaini.

96 Hussaini was charged with one count of unlawful detention, which is alleged to have occurred on 29 September 2004, some time after noon, towards 12.30. The State alleged that, when the complainant was out walking her dog, she saw a white "courier style" van. The driver stopped and spoke to the complainant. He spoke softly and she had some difficulty in hearing him. He asked her for the name of the street. She approached to tell him. As she came up to the window, he lent across and, with his left hand he grabbed her tightly by her left hand or forearm, so that she could not get away. He said words to the effect of, "Get in the car, I'll give you 20 bucks for an hour". He was still speaking quite softly. After a short struggle, she managed to pull away. She went quickly home.

97 The complainant described the van as being a "tall white van". She described a kangaroo on the side of the van, towards the front near the driver's window. She thought it might have been yellow. She described gold and green stripes, running along the side of the van and at the back. Her impression was that it may have been a Mercedes van "because of the logo".

98 The complainant gave a description of the person as looking like a "part-Asian" male, in his late 20s to mid-30s, with tanned skin, reddish-brown hair partly covering the ears and a bushy moustache. She participated in the making-up of an identikit picture, which was widely publicised.

99 The appellant had left work on the day in question at about 11.21 am. It appears to be common ground that had he desired to do so, he could have driven to the location where the offence occurred by 12.30 or thereabouts. His father-in-law had a business which had involved a number of vans which bore a green kangaroo logo near the door and gold and green stripes. By the date of the alleged offence, the appellant's father-in-law had apparently sold the business, and had retained one Toyota van, to which the appellant had access.

100 There was evidence from a neighbour of the appellant who said that in October of 2004, when he read a description of the van used in the offence, he believed that it fitted the van which he had seen being used by his neighbour, the appellant. He also was of the view that the identikit picture which had been published in the newspaper bore a "striking resemblance" to his neighbour, the appellant. His Honour told the jury that this witness' opinion was not evidence of identification.

101 The complainant identified the appellant from a digiboard which she was shown on 25 November 2004. She had on two other occasions been shown digiboards which did not contain an image of the appellant, and she made no identification from either of them.

102 The appellant gave evidence which was to the effect that he had left work early on the day in question, but had gone straight home because of an injury to his shoulder. He was "confused" about what car he had driven at the time. It may have been a green Daihatsu Charade, or it may have been the van which belonged to his father-in-law. He gave evidence, which was supported by a number of witnesses to some degree, to the effect that he had at one time injured his left arm, or left shoulder. His evidence was that the injury was so severe that he would not have been able to hang on to the complainant in the manner she described. He said he did not own a white polo shirt of the type described by the complainant. Evidence was adduced of his good character.

103 The appellant's father-in-law produced photographs of vans said to be vans owned by the company. The State's submission to the jury was that it appeared that a person or persons unknown may have "doctored" the photographs, which were digital photographs. The precise significance of that is not easy to ascertain, in the absence of a transcript of the addresses of counsel. However, we understand from the submissions made by the appellant's counsel before us, that there may have been evidence adduced on behalf of the appellant, in relation to the van, which was in some respects implausible. It may also have been the case that the jury could have considered that the appellant's shoulder injury was, having regard to the evidence of some witnesses, not quite the disability which he portrayed. Those matters may be of significance in relation to the "Liberato" issue. We deal with that issue later.

104 So far as the issue concerning the compilation of the digiboards is concerned, there were, as we have noted, three digiboards. Two were viewed by the complainant on 19 October. They each contained 12 photographs. A significant number of the persons in those photographs, although not all, had either significantly or slightly almond-shaped eyes, and a number would appear to have been recognisable by most people as being of Asian origin. A number had oval faces or broad foreheads and a number had thick, dark hair. In that respect, they resembled the identikit compiled by the complainant. However, only a very few had any kind of facial hair, and that facial hair was not luxuriant. In that respect, they did not match the thick moustache which was shown on the identikit photograph.

105 However, on 25 November, the complainant was shown a further digiboard containing 12 photographs. Each of the persons on that digiboard had a moustache, and many had long, or thick, head hair.

106 It is submitted on behalf of the appellant, that most of the digiboard images do not look "Asian". However, the complainant's description was of a man who appeared to be "part-Asian", and her identikit does not resemble what most people would imagine as a stereotypically Asian face. The only "Asian" features on the identikit are the very slightly almond-shaped brown eyes and the almost straight dark brown hair. Most, but not all of the photographs on the digiboard, are of persons with almond-shaped, or slightly almond-shaped, eyes.

107 The appellant's counsel suggested to us that the way in which the digiboard was compiled meant that the appellant stood out, since his was the only image which had reddish-coloured hair. The discussion between the bench and counsel reveals the difficulty of characterising the appellant's image in this way. One of us referred to the appellant's hair as being "purplish" and another described it as having a "blue" tinge. One might speculate that it had been altered so as to attempt to remove a reddish tinge, because it had not been found possible to compile a set of 12 images of persons of vaguely Asian appearance who also had a reddish tinge to their hair. The question was not explored at trial, because no issue was taken with the compilation of the digiboard, by experienced defence counsel. However, whatever may explain the colouring, the appellant's photograph is not the only one in which the hair has an unusual tinge. Numbers 9 and 12 also have a faintly blue or blue-green colour to their hair (as do some of the others, although much more faintly, when examined in good light). Number 3 has a reddish tinge to his hair, although very faintly. A number of the other photographs have brownish, rather than very black hair.

108 In our view, it could not be said that the appellant's image stood out so distinctly from the others on the digiboard that the complainant must inevitably have identified him, or in a way which would lead a person who was unsure of their identification to have their attention attracted particularly to him. We are fortified in that conclusion by the fact that this question was not apparently one which occurred to trial counsel for the appellant. He cross-examined the complainant about the digiboard, and in particular about the goatee which the appellant is shown to have on that digiboard. He cross-examined the investigating officer briefly about the composition of the digiboards, but did not touch on this issue. This aspect of the ground is not made out.

109 Turning to the question of delay, his Honour did not direct the jury in relation to delay. So far as delay is concerned, it appears that research and common sense coincide in suggesting that memory for an event declines quite quickly after it. Once the initial period of decline has passed, it appears that memory continues to decline to some degree, although the rate at which it does so is not clear, and may vary depending upon a number of factors. The conclusion of Brewer et al is that the "retention interval" (which we understand to refer to the time between observation and identification) would be expected to interact with a number of other variables. Without knowing the parameters of those variables, they suggest that "debates about the specific time course and the extent of any performance decline with retention interval are unsatisfying. This is not to say that retention interval effects are unimportant; rather, we question whether it is particularly meaningful to try to say much more than that identification performance is likely to decline with retention interval" N Brewer et al, "Eyewitness Identification" in N Brewer and F Williams (eds) Psychology and Law: An Empirical Perspective (2005) 177, 192. In our view, since delay does have an effect upon the accuracy of identification, even if it is not possible to say anything about the likely effect of any particular period of delay, trial judges should warn juries that delay between observation and identification is a factor which they should consider in assessing the accuracy of an identification. His Honour did not do so in this case.

110 His Honour having not given what we would regard as a desirable component of the identification warning, a question arises as to whether that failure is significant enough to lead to the conclusion that the warning was, as a whole, inadequate. It is desirable, therefore, to set out what his Honour did say. His Honour said this:

Now, the state's case rests, of course, to a very significant extent on the identification of the accused person made by the complainant in this matter, Courtney Amanda Melville when she picked out photograph number 2 in the third photoboard, if you are satisfied that that's what occurred. As I said, all matters of evidence are matters for you not me. The significance of that is, members of the jury, that where the state's case rests on evidence of identification of an accused person by a state witness, there is a special need for caution before convicting.

The reason for that is that in the past, such evidence has, on occasion, been found to be mistaken later on. That's in the experience of the law and, of course, evidence of that kind very often is given by people who are completely honest witnesses who are simply mistaken but because they are honest witnesses their evidence tends to be accepted by juries. A statement, 'That's the person I saw' or, 'Who did something to me', can never, of course, be more than an expression of opinion, in other words, 'In my opinion that is the person who I saw and who did something to me.'

With that in mind, members of the jury, it's necessary that I warn you that whatever degree of confidence is exhibited in an identification it is necessary for you to scrutinise the evidence of identification with great care before acting on it. Many times, of course, such evidence can be entirely accurate and therefore reliable. The difficulty is to say when that is the case.

There are rules, of course, of commonsense which any of us would be able to work up if necessary in relation to the issue of identification including, was the person identified known to the identifier? What opportunity for accurate observation existed? What was the light like? How much time existed for the observation? What aspect of the person was presented to the identifier and how much of the body of the person was presented? Were movements observed?

At what distance did the observation occur? Were there intervening or other distracting movements? Were there any particular factors which might have affected the situation? Were there any apparent or obvious peculiarities of the person identified which might have assisted? Was the witness under a state of stress and, if so, what effect would that have been likely to have had? Some people, of course, under stress have increased powers of observation, others would have their powers of observation lessened. What is likely to have happened here?

Did the witness have reason to try and observe and remember the person? What was the description given? When was the description given? How was the identification made? So matters of that kind are all to be taken into account, members of the jury. Some of them, of course, would be of less relevance than others in this particular case.

111 It is perhaps important to note that his Honour did not purport to give the jury an exhaustive list of relevant factors. Rather, he said that "matters of that kind" were all to be taken into account. Having given this warning which identified the core problem with identification evidence, and which referred to a considerable number of relevant factors, his Honour went on to pick out what he described as "some aspects" of the complainant's evidence which would have the potential to affect the jury's view of her evidence. They included the description she initially gave, which his Honour summarised, the reference to the white shirt, the brevity of the identification (a maximum of five minutes), the degree to which she saw the offender's face "full on", and the fact that there was a struggle. His Honour also told the jury that identification by photoboard was not ideal, being two-dimensional and "just a snapshot". He reminded the jury that they could watch the video of the identification and then went on to tell the jury that they would need to consider the complainant's evidence in relation to the type of vehicle also, dealing with that issue in some detail.

112 The answer to the question of whether the absence of a reference to delay leads to a conclusion that the warning, as a whole, was inadequate, is, we think, to be found in the observation in Domican, at 562, that " ... the trial judge should isolate and identify for the benefit of the jury any matter of significance which may reasonably be regarded as undermining the reliability of the identification evidence" (emphasis supplied). We would understand their Honours, in the reference to "any matter of significance" to be distinguishing matters which might have a real bearing, in practical terms, upon the jury's assessment of the identification evidence. What those matters may be will vary from case to case. The words "of significance" would, we think, exclude reference to matters which would have a remote or hypothetical effect only upon the assessment by the jury in any particular case. Indeed, given the purpose of the warning, for a judge to put the weight of his or her authority behind a warning which referred to arguably insignificant matters would run the risk of confusing a jury and perhaps leading members of the jury to give inadequate attention to those matters which were truly "of significance". We note that in Festa at [74] per McHugh J, [175] per Kirby J and [218] - [219] per Hayne J, there are suggestions, expressed in different ways, that the effect of a failure to warn will depend upon the significance of the matter omitted, even if it is a matter which may reasonably have some effect upon the jury's assessment.

113 In Domican's case, delay in identification was obviously regarded by the High Court as a relevant matter. However, that was in circumstances where nearly nine months had elapsed before the relevant witness identified the appellant, by which time not only was the appellant a definite suspect, but the witness had seen him on television on a number of occasions and also allegedly in the vicinity of her home (at 562). In the present case, while the period of delay before the complainant identified the appellant was referred to once by counsel for the appellant in cross-examination of her, the more significant factor appears to have been the period which elapsed between the observation of the offender and the compilation of the identikit image. The identikit was compiled on Friday, 1 October, only two days after the incident in question (blue/green AB 41). It is to be noted that the appellant's counsel at trial apparently did not consider the delay to be of such significance as to call for a warning from his Honour, since he did not raise that at the conclusion of his Honour's direction (ts 209). The attitude of trial counsel is, of course, not decisive, but it is relevant.

114 The question of whether the delay in this case was a "matter of significance" which might reasonably be regarded as undermining the reliability of the complainant's identification, is a matter about which, having regard only to the printed transcript, minds might reasonably differ. It is, in our view, appropriate in those circumstances for this court to defer to the assessment of his Honour the learned trial judge, who had the advantage of seeing the complainant give evidence and of viewing the videotape of the identification she performed. It is apparent from his Honour's direction on the topic that he had given consideration to the question of what factors he should make the subject of his warning to the jury, and we are not able to say, in the particular circumstances of this case, that he erred in omitting reference to delay.

115 Finally, there was a question raised by the State about whether his Honour should have directed the jury to consider whether the complainant had made a positive identification at all. What she said when initially viewing the digiboard was "number 2 looks quite familiar" (that being the appellant). She clarified that he was familiar in relation to the offence in question. The police officer then said to her, "That's a fairly quick identification. How certain ...?" She said, "I just recognise the hair, the eyes and just the facial expression. It looks quite, yeah, but I can't say that's definitely him one hundred per cent but it does look like him from what I can remember" (blue/green AB 177).

116 Had that remained the only evidence of identification, there might have arisen a question of the kind considered by the High Court in Pitkin v The Queen [1995] HCA 30; (1995) 69 ALJR 612. In that case, the witness had said of photographs, "This looks like the person". The court held that if there existed "verbal ambiguity" in relation to the words used by a witness at the time of identification, then, "without further explanation or elucidation", the accused must have the benefit of any ambiguity (at 615). However, in this case, there was elucidation and explanation from the complainant. In her cross-examination, when asked to agree that she had not originally told the police that the man had a goatee, her response was, "No, I didn't but just as soon as I saw this picture I knew it was him straight away [referring to the digiboard]" (blue/green AB 31). In re-examination, she explained what it was about the photograph that she recognised. Counsel put to her that she could not be "100 per cent" that number 2 was the offender. She agreed that that was correct, but went on to say that she did not mean it in terms of "I'm not too sure if that looks like him or not" (blue/green AB 34). Although what she then said was not entirely clear, it appears that what she was trying to convey was simply that she could not say "that's absolutely him" (blue/green AB 35), but she was not saying that she was not able to determine whether it was the person.

117 This evidence does raise however, a question of whether there was a factor which might have had significance in relation to the jury's assessment of the complainant's identification. If she had appeared to have had real doubt at an earlier stage, but had become more confident, that would be a matter about which it would have been appropriate for the judge to have directed the jury. Confidence immediately following the identification may be relevant to accuracy, from the materials before us, but a witness' confidence may be influenced by later events, and even by the experience of being called as a witness (because this suggests that the person had something relevant to say): G Wells et al, "Eyewitness Identification Procedures: Recommendations for Lineups and Photospreads" (1998) 22(6) Law and Human Behaviour 1, 19; G Wells et al, "Eyewitness Evidence: Improving Its Probative Value" (2006) 7(2) Psychological Science in the Public Interest 45, 66; G Wells and E Olson, "Eyewitness Testimony" (2003) 54 Annual Review Psychology 277, 284.

118 However, whether there was a possibility that a witness' confidence had increased subsequent to the identification, would depend upon an assessment of what was said at that time, the way in which it was said, and what the witness had said subsequently. We have before us a record of what the complainant said, but not of the way in which she said it. Her evidence at trial seems really to have been to the effect that she was identifying the appellant when she saw the digiboard, but that she said then (and we think maintained at trial) that she could not be "absolute" about that. It may not have seemed to any of those present at the trial that there was any relevant alteration in her confidence and, in the light of her evidence, it would have been clear that it was her view that what she had said at the time of viewing the digiboard was to be regarded as an identification.

119 When the State raised the question of whether this should have been the subject of direction, its absence was embraced by counsel for the appellant as a further defect in his Honour's direction. It is our view however, that his Honour was not required to direct the jury either to assess whether there had been an identification at all, or about any possible variations in the complainant's confidence. The fact of identification does not seem to have been left in doubt. Whether there was a fluctuation in confidence is, so far as we are concerned, speculative. Further, had his Honour referred to confidence, he would have been as likely to disadvantage the appellant as to assist his case. A direction of that kind might have only served to reinforce a view, open on the evidence of the complainant at trial, that she was merely cautiously and consciously adverting to the difficulties inherent in identifying a person not known to her previously.

Conclusions concerning identification: summary

120 For convenience, we list briefly here the main conclusions about issues which have arisen in these two appeals concerning certain suggested aspects of the identification warning.

(a) A trial judge is not required to direct a jury that digiboard identification is unreliable and dangerous per se, or that it is inferior to other types of identification. [55]

(b) A trial judge should usually direct a jury that identification from a digiboard may be affected by the fact that the images on the digiboard are static and two-dimensional. [48]

(c) In some cases, it may be necessary for a trial judge to direct a jury that identification may be affected by the fact that images on a digiboard do not include identifying (or exclusionary) factors such as height, build and posture. [47] - [48]

(d) In some cases, it may be necessary for a trial judge to direct a jury that they should take into account the risk that a relative identification has occurred, rather than actual identification. [62]

(e) A trial judge is not, as a general rule, required to direct a jury that discrepancies between a description given by a witness and the appearance of the person identified by the witness may suggest that the identification is unreliable. [68] - [70]

(f) A trial judge is not required as a general rule to direct a jury that "stress" is a factor that may make a witness' identification suspect. [87]

(g) Where there is any delay between observation and identification, a trial judge should direct a jury that delay is a factor to be considered in assessing the accuracy of identification. [109]

121 The "Liberato" issue (Liberato v The Queen [1985] HCA 66; (1985) 159 CLR 507) can be fairly shortly dealt with. That is a direction which is or may be required in some cases where there is a conflict between the evidence of the defence and prosecution witnesses. The appellant Hussaini submits that "[o]ne of the central issues for resolution by the jury was whether to believe the version of events given by the [c]omplainant or the [a]ccused" (amended white AB 14). It will be apparent, from our outline of the trial, that this was not one of the central issues at all. The appellant did not directly attack any aspect of the complainant's "version of events". Neither he nor his witnesses had anything to say about the question of whether the complainant had been out walking as she described, had been accosted by a man as she described, and had run away as she described. Rather, his case was that, whatever may have happened to the complainant, she had been mistaken in her identification of him and that, for various other reasons given by him and supported to an extent by others, he could not have committed the offence.

122 During the course of the appeal, the oral submissions of counsel put the need for such a direction on a slightly different basis. It was suggested that there was an aspect of the appellant's case, being the question of the photographs of the vehicles, in relation to which the jury would be entitled to find that the account put forward by the appellant and his witnesses was "rubbish". It was suggested that, in that case, a "Liberato" style direction was necessary.

123 However, his Honour did deal very clearly with the question of what the members of the jury should do if they found that something said by the appellant, or evidence proffered by defence witnesses, appeared to be untruthful. In relation to the suggestion made by the prosecutor that the appellant had told untruths both in what he said to the police officers and in the course of his evidence at trial, his Honour told the jury that, as a matter of common sense, that would be relevant to his credibility, but went on to say that:

... it would not be evidence which would tend to support an ultimate inference of guilt, in other words, you would be able to look at that when you came to consider his credibility but you would not be able to use that as evidence directly showing or tending to show that he was guilty of the offence. (blue/green AB 168)

124 His Honour went on to explain that there might be a variety of reasons why people told lies, including to bolster a just cause.

125 In relation to the photographs of the van, his Honour referred to the prosecutor's submission that there had been apparently some alteration of one or more photographs. He told the jury that they were entitled to scrutinise those photographs, but that they should refrain from speculating about matters in respect of which there was no evidence. If they thought that there had been some human agency involved in altering those photographs, his Honour said "that of course could not be evidence of guilt of the offence on the part of the accused person". He went on to say that even if an alteration had occurred, there was no evidence that the accused had been involved in it. Even if the accused had deliberately lied about the appearance of the van, that would be relevant only to his credibility (blue/green AB 167 - 170).

126 Finally, in relation to this issue, it is submitted that the following statement made by his Honour could have confused the jury and led them to believe that conviction or acquittal was to be determined by favouring one party's evidence over the other:

Whether or not the state has satisfied you on the whole of the evidence that the person who detained the complainant on that day, if you are satisfied as to that, was the accused person - and of course for you to be satisfied as to that, it would be necessary for you to be satisfied that the accused person's denial was false and to be satisfied that you had to reject his evidence as well as being satisfied that the complainant's evidence was both truthful and accurate, that is reliable.

127 This is a difficult sentence to follow, on the page. It is often the case, in relation to sentences of this length and complexity, that they are spoken in a way that untangles the various elements and emphasises some rather than others. However, even on the page, it is clear, reading the quoted portion with the preceding page, that his Honour was identifying the central issue for the jury as being whether or not the State had satisfied them on the whole of the evidence that the person who detained the complainant was the accused person. That was the central issue, and in articulating it, his Honour was clear about the burden of proof. His Honour was also in that sentence indicating that, for them to be satisfied that it had been the accused, it would be necessary for them to be satisfied that the accused person's denial was false. That is also plainly correct, since his evidence, if not positively rejected, was capable of giving rise to a reasonable doubt. His Honour told the jury in that passage that "as well as" rejecting the appellant's evidence, it would be necessary for them to be satisfied that the complainant's evidence was both truthful and accurate before they could convict. That, too, is correct. It does not appear to us that this passage would have left the jury in any doubt about the way in which they were to approach their task.

128 Leave has not been given in relation to the ground of appeal raising the Liberato question. We would not grant leave in respect of it. We would dismiss Hussaini's appeal.

129 For the reasons given above, it is our view that both appeals must be dismissed.