The State of Western Australia v Roe - [2017] WASC 124

JURISDICTION : SUPREME COURT OF WESTERN AUSTRALIA
IN CRIMINAL

CITATION: THE STATE OF WESTERN AUSTRALIA -v- ROE [2017] WASC 124

CORAM: BANKS-SMITH J

HEARD: 18 APRIL 2017

DELIVERED : 3 MAY 2017

FILE NO/S: INS 18 of 2017

BETWEEN: THE STATE OF WESTERN AUSTRALIA
Prosecution
AND
PHILLIP JOHN ROE
Accused

Catchwords:

Criminal law - Evidence - Facebook identification - Warnings and reliability

Legislation:

Nil

Result:

Evidence admissible

Category: B

Representation:

Counsel:

Prosecution : Mr D T Carlson
Accused: Mr D C Rice

Solicitors:

Prosecution : Director of Public Prosecutions (WA)
Accused: Griffiths Rice & Co

Case(s) referred to in judgment(s):

Alexander v The Queen [1981] HCA 17; (1981) 145 CLR 395
Bayley v The Queen [2016] VSCA 160
Collard v The Queen [2000] WASCA 417
Dia v The Queen [2014] NSWCCA 9
Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555
Peterson v The Queen [2014] VSCA 111
R v Crawford [2015] SASCFC 112; (2015) 123 SASR 353
R v Janissen [2013] QCA 279
R v Kearney [2013] SASC 121
Strauss v Police [2013] SASC 3; (2013) 115 SASR 90
Winmar v The State of Western Australia [2007] WASCA 244; (2007) 35 WAR 159
BANKS-SMITH J:

The issue

Mr Roe and a co‑accused (E) are charged with one count of criminal damage by fire. It is alleged they set fire to the complainant's house. The complainant saw the offenders and was later told that E was one of them. The complainant searched E's profile on Facebook, found a photo of Mr Roe and told police that Mr Roe was the other offender.
Identity is in issue. Whilst the prosecution will rely on other circumstantial evidence, it seeks to lead evidence as to the complainant's identification of Mr Roe through Facebook. The State seeks a direction before trial as to the admissibility of that evidence. Mr Roe seeks its exclusion.

The State's case

The following facts are taken from the State's Statement of Material Facts and statements made by the complainant and the police officer involved in the identification process.
On 8 August 2016 the complainant was outside his unit in Broome. He heard a neighbour, SA, talking on her phone. The complainant was worried so he went inside his unit and locked all doors and windows. It was dusk and dark outside. Outside lights were turned on.
The complainant saw SA leave in her car and return ten minutes later. When she returned, she was in the front passenger seat, a male was driving and another male was in the rear seat.
The complainant states that he could see the two males clearly because the car was driven up to his front door and the front lights of the unit were on.
The complainant did not know the men. He described the driver as a 'big bloke, Aboriginal' who was 'wearing a white bandana with black on it'. He described the male in the back seat as 'Aboriginal and looked a bit smaller' than the driver. Both accused were wearing shoes but the complainant could not recall what the offenders were wearing.
The complainant stood to the side of his pool table near the front door. He had a clear view of the front window and saw the two males come to his front door and start kicking it.
The two males went to the side of the unit, entered the patio area and began kicking the glass part of a sliding door until it shattered. The men then set fire to the patio area of the unit. The complainant could still see the males from where he was standing.
After the fire, the complainant spoke to one of his neighbours, P, who saw the offenders and told him E was one of the males and she had known E since he was young.
After this conversation the complainant did a Facebook search for E (by his full name). He did this two days after the fire. On E's Facebook page, he located a photo of E and another person (Mr Roe) sitting together on a fence. He also saw from E's Facebook page that they were Facebook friends. Mr Roe used the name 'Fillip Ro' on Facebook.
The complainant went to the police station and told them he had identified the people who set fire to his house through Facebook. He logged on to his account and showed the police the photos he had found on Facebook. The police printed the photo of Mr Roe and E together and separate photos of each of E and Mr Roe from their profile pages. The photo of Mr Roe and E together on the fence and the separate photo of Mr Roe are relied upon by the State.
The separate photo of Mr Roe is an almost full length shot and shows him standing in front of a large rock wearing jeans and a baseball cap and wearing sunglasses. He is not wearing a shirt and has many tattoos on his arms and torso.
The police had intended to conduct a digiboard viewing with the complainant but after the Facebook photos were obtained, they did not proceed with the viewing.
Mr Roe accepts that the two photos said to be photos of him which were printed from Facebook are in fact photos of him. He has confirmed he used the name 'Fillip Ro' on Facebook.
Mr Roe denies any involvement in the offence.
In short, the State says that the probative value of the Facebook identification evidence outweighs any prejudice in admitting the evidence. Directions to the jury by the trial judge as to the dangers of identification evidence will overcome any prejudice. The weight to be given to the evidence is properly a matter for the jury.
The defence accepts that identification of suspects by Facebook can be admissible if there are not clear dangers in doing so that would lead to unfair prejudice. It says that in this case there are dangers that cannot be overcome and that the court should exercise its overriding discretion to exclude the evidence.

Facebook as an investigative tool - general

The exponential increase in the use of social media has led to 'a new generation of private investigators'.[1] It is not uncommon for victims and witnesses to search Facebook looking for information about an offender, or a person they may think is an offender.[2]
[1] Strauss v Police [2013] SASC 3; (2013) 115 SASR 90 [34].
[2] Straussv Police [35].
There is a growing body of case law dealing with the issue, although as yet, the question of the reliability of Facebook identification and the nature of any directions to a jury has not been considered in published reasons of this court. Before turning to cases in other jurisdictions, I will briefly refer to the general position in this State as to identification evidence.

The discretion to exclude identification evidence

Traditionally identification parades were said to be preferred to photoboard identifications. Concerns about photoboard identification were expressed in particular in Alexander v The Queen.[3]
[3] Alexander v The Queen [1981] HCA 17; (1981) 145 CLR 395.
In Winmar v The State of Western Australia,[4] the Court of Appeal considered the reservations as to the reliability of photoboards identified in Alexander. In summary, those reservations include:
(a)the absence of the accused from the process limited the ability of the accused to test the veracity of the process;
(b)photographs may appear to be 'mug shots' and may give rise to an assumption that the police would only take a photo of the accused if they were a suspect;
(c)an inference may arise from the fact the police have a photo of the accused, particularly if identification occurs prior to being questioned or an arrest;
(d)the displacement effect, particularly where a person first identifies a suspect from a photograph and then later purports to identify them at an identification parade; and
(e)the static and black and white character of old style photoboards.
[4] Winmar v The State of Western Australia [2007] WASCA 244; (2007) 35 WAR 159 [39] ‑ [44].
Regardless of those concerns, it was held in Alexander that evidence of out of court identification from photoboard identifications was admissible, subject to the trial judge exercising a discretion to exclude it.[5]
[5] Alexander v The Queen (399); see also Collard v The Queen [2000] WASCA 417 [71].
In Winmar, the Court of Appeal then considered the position with respect to the development and use of digiboards for identification. At that time, there was no other authority that dealt with digiboard identification. The Court of Appeal explained the digiboard identification process and its disadvantages. For example, an issue with the digiboard process is its two‑dimensional nature and that part only of a person is seen (the head and not the whole body).[6]
[6] Winmar [49].
Circumstances where the displacement effect could arise were no more likely to occur than with a photoboard.[7]
[7] Winmar [44].
A digiboard process may have advantages over an identification parade. In an identification parade, a suspect's demeanour may be different because, for example, regular fillers for such line-ups may appear quite relaxed.[8]
[8] Winmar [51].
Importantly, the Court of Appeal concluded that:[9]
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.
[9] Winmar [55].
The Court of Appeal then provided guidance as to aspects of the identification warning that might be given to a jury where there has been digiboard identification, depending on the facts.[10]
[10] Winmar [120].
The history of the need for a warning by way of direction and when such warnings may be required are also discussed more generally in Winmar.[11] Relevantly, the Court of Appeal identified some core warnings as to the dangers of convicting on disputed identification evidence:[12]
(a)in every case, there is a danger that an honest witness may be mistaken, and that an honest but mistaken witness may be convincing;
(b)in most cases, there may be to a greater or lesser degree a danger in that identification evidence may be open to suggestion; and
(c)in most cases, there may be to a greater or lesser degree a danger that the witness will not have had an adequate opportunity to observe the offender so as to identify them.
[11] Winmar [4], [7] ‑ [26]; Domican v The Queen [1992] HCA 13; (1992) 173 CLR 555, 561 ‑ 562.
[12] Winmar [12] ‑ [15].
The Court of Appeal also noted that the issue of a lack of a precise description does not mean that a witness is not able to recognise someone:[13]
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.
[13] Winmar [66].
Further, discrepancies in descriptions do not of themselves lead to the need for any direction: discrepancies may not be of significance.[14]
[14] Winmar [68] ‑ [70].
In summary, the approach of the courts in this State with respect to photo identification whether by photoboard or digiboard is that it is admissible. It can be excluded in the discretion of the trial judge. Appropriate directions should be given to the jury if there is a specific weakness as to unreliability of identification evidence.
An assessment of the position with respect to Facebook identification in this State must take place against the backdrop of Winmar. However, the line of authorities in other States provides guidance.

Facebook identification - other jurisdictions

In Strauss v Police, a man was beaten after a work function. A colleague witnessed the attack. She assumed a particular woman was the reason for the attack. Later that night she searched the woman's Facebook page and found a photo of the defendant (who was named) who she said looked like the offender. The witness gave the name of the defendant to the victim. The victim then searched the defendant's name and identified him as the offender. The defendant was convicted and appealed. The identification evidence was only one aspect of the appeal.
Peek J considered in some detail the practical manner in which Facebook operates and the risks as to the accuracy of photos that are posted or tagged on Facebook, usually without any auditing process.[15] Peek J considered the risks associated with Facebook identification are more pronounced where the victim is given the name of the alleged offender and so searches for the person by that name, their identification perhaps a product of hours of staring at the person's photographs. The image would potentially displace a memory the victim has of the person, particularly where the person's memory of the event is otherwise affected (eg by alcohol).
[15] Strauss v Police [12] ‑ [37].
His Honour upheld the appeal, noting a number of reasons why the particular Facebook identification was unreliable: the identification was not spontaneous but studied;[16] officers had not taken a description of the offender at the scene;[17] the victim was drunk at the time of the attack and prevailing light conditions were not good;[18] there was no formal identification process conducted.[19]
[16] Strauss v Police [115].
[17] Strauss v Police [46].
[18] Strauss v Police [76].
[19] Strauss v Police [58] - [65].
In Dia v The Queen,[20] the charge was one of armed robbery. A group of men broke into a house and terrorised the victims. One of the victims, Ms McCann, knew one of the offenders, Mr Fawaz. A detective and Ms McCann undertook a search of Facebook for Mr Fawaz's page. She identified Mr Fawaz in a photo on the page of a group of men. The following day, Ms McCann went onto Mr Fawaz's Facebook page and went through his photos to see if she could identify any co‑accused. She scrolled and came across the profile of a Mr Khawaja and she recognised him as the person who had stabbed her brother. The applicant (who went by the names of Khawaja and Dia) contended at trial that there were defects in the identification evidence. In particular, it was contended that Ms McCann went on Facebook with an expectation that one of Mr Fawaz's friends was involved in the offence. There were also inconsistencies in the physical descriptions of the offender given by Ms McCann (different height, hair length). She had been 'stoned' at the time. The trial judge gave directions about these 'defects'. The applicant was convicted and appealed.
[20] Dia v The Queen [2014] NSWCCA 9.
The Court of Appeal considered the 'defects' had been the subject of proper directions at trial and dismissed the appeal. As to the criticism of the Facebook identification process, the court said:[21]
The criticism of the circumstances in which Ms McCann made her identification of the applicant while available to be raised, is not decisive. … Moreover, Ms McCann was not looking at those images with the intent of identifying other persons involved in the offences but to confirm her identification of Mr Fawaz.
It is clear that on the day following the offences, Ms McCann's intent when examining the various Facebook sites was different. She freely accepted that she did so in order to 'identify' if she could, any of the other persons who had participated in the offences on the night. When doing so she was able to identify the applicant. There are obvious dangers associated with any identification from a photograph and there were some particular dangers associated with this mode of identification. These matters were not only put to Ms McCann in cross-examination but clear warnings as to the reliability of the identification were given by his Honour to the jury.
While there are dangers in such a process of identification, it is also a method which is acceptable and evidence concerning such identification is admissible. The process of identification undertaken by Ms McCann was analogous with her attending a police station and examining a photographic array or going to a police station to view a line-up.
[21] Dia [66] ‑ [68].
In Peterson v The Queen,[22] the victim was attacked by two men over a drug debt. One of the offenders stabbed him with a knife. The victim gave a description of that person to the police. A friend visited the victim in hospital and mentioned the applicant's name. The victim looked at the applicant's Facebook page and believed it was the person who had stabbed him and identified him to the police.
[22] Peterson v The Queen [2014] VSCA 111.
The police did not attempt to have the victim formally identify the man who stabbed him by means of a photoboard or identification parade. Two reasons were given for declining to do so. The applicant looked similar in the Facebook photo to photos the police had which may suggest the victim would too easily identify the applicant. Further, there was concern about displacement: that the victim would select the applicant from the photoboard, not because he resembled the offender but because he resembled the person identified in the Facebook photo. Identification from, for example, a photoboard, 'would be open to the criticism that it was polluted by the antecedent viewing of the photograph on Facebook'.[23]
[23] Peterson [47].
Prior to trial, the issue of admissibility of the identification issue was the subject of an interlocutory ruling. The trial judge refused to exclude the identification evidence. That ruling was then the subject of appeal.
In contrast to comments in Strauss v Police,[24] the Court of Appeal considered there were sound reasons to refrain from a formal identification process after the Facebook photo had been viewed.[25] It considered the evidence was clearly probative and so the question was whether the frailties in the evidence gave rise to a danger of unfair prejudice. Those frailties were said to include that the victim had given only a limited description to the police after the offence; issues as to the circumstances of his initial observation of the offender; and the suggestibility involved in searching Facebook for the named person.[26]
[24] Strauss v Police [65].
[25] Peterson [47] ‑ [48].
[26] Peterson [53] ‑ [54].
The ruling refusing to exclude the evidence was upheld. The Court of Appeal held that the frailties in the evidence and the criticisms made of it, are matters which the jury are capable of evaluating with the benefit of proper judicial direction. The reliability of the identification from the Facebook page is pre-eminently a jury question.[27]
[27] Peterson [55] ‑ [56].
In R v Crawford,[28] The victim was given the name of the alleged offender by the police. The victim then searched Facebook on his phone for profiles with that name. He looked at the top five or six profiles on the results page, and recognised the defendant. About two months later, the police conducted a photo identification session where the victim was provided with 12 photos. He identified the defendant from those photos.
[28] R v Crawford [2015] SASCFC 112; (2015) 123 SASR 353.
At trial, the defendant challenged the admissibility of the identification evidence on a voir dire, relying on concerns identified in Strauss v The Police. The trial judge admitted the evidence and gave extensive directions to the jury, including as to possible contamination of the photographic identification because of the Facebook identification and the displacement effect. The appeal was unsuccessful. The Court of Appeal was critical of the police conduct in providing the name of the alleged offender to the victim, particularly as it was likely the victim would conduct a Facebook search and so potentially compromise the identification process.[29] However, the majority held that the evidence had probative value (even if slight) and any prejudice that might otherwise arise from it being given more weight than it should was dealt with by the appropriate and strong directions given to the jury.[30]
[29] Crawford [10] (Gray J), [54] (Peek J), [127] (Nicholson J).
[30] Crawford [29] ‑ [30], [36] (Gray J), [165] (Nicholson J), (Peek J in dissent).

In Bayley v The Queen,[31] the Court of Appeal overturned a conviction for rape in circumstances where the case on identity relied entirely on the alleged victim's identification of the appellant from a photo of him on Facebook. At the time of the identification, the appellant was the high profile suspect in the murder of Jill Meagher. The alleged victim was looking at stories about Jill Meagher on Facebook when she saw the photo of the appellant and so was aware he was a suspect. The identification by the alleged victim of the rape occurred some twelve years after the event. The court stated that the identification was:[32]
[I]n some respects no better than dock identification. Indeed, it could reasonably be viewed as worse. She was faced with a single photograph in circumstances that were themselves highly suggestive of guilt of the most serious crimes imaginable.
[31] Bayley v The Queen [2016] VSCA 160.
[32] Bayley [94].
The court considered that any probative value that the evidence theoretically had was outweighed by the risk to the applicant of unfair prejudice.

Summary of authorities

I take into account that some of the above cases consider applications to exclude evidence in the context of various statutory provisions that do not apply in this State.[33] However, the authorities are useful in that they identify the potential dangers of Facebook identification.[34]
[33] In particular, Evidence Act 2008 (Vic) s 137; Evidence Act 1995 (NSW) s 114.
[34] There are other cases referring to Facebook identification but they are of limited assistance on this application: eg R v Kearney [2013] SASC 121; R v Janissen [2013] QCA 279.
There is a uniform theme that the risks associated with Facebook identification can generally be ameliorated, except in extreme circumstances such as in Bayley, by proper directions to the jury.

Submissions of the parties

In this case, Mr Roe submits that the Facebook identification evidence is unfairly prejudicial and should be excluded because:
(a)the complainant was given the name of the alleged co-accused and searched for that particular name on Facebook;
(b)the complainant assumed having seen a photo on E's Facebook page of E and Mr Roe together that Mr Roe was the other offender;
(c)the police did not conduct a photoboard or line up;
(d)the complainant did not give police a clear description of the offender after the offence or identify any distinguishing features;
(e)the complainant did not identify the men on Facebook for some two to three days after the offence; and
(f)the Facebook photo of Mr Roe appears to be 'substantially different' to the photo of Mr Roe taken when he was arrested, 10 days after the offence.
The State relies on the following:
(a)the complainant was not given Mr Roe's name and so did not direct his attention to a particular person when identifying Mr Roe. Similarly to the facts in Dia, the complainant in this case knew the name of the co-accused and then used that name to look at other photos from which he identified Mr Roe. In Dia the evidence was admissible;
(b)the identification process was of more probative value than in Crawford and Peterson, and in those cases the evidence was not excluded. In Crawford and Peterson the complainant was given the name of the alleged offender and searched by that name;
(c)in searching through Facebook, the complainant had the opportunity to see more than one photo (in contrast to Bayley). The process was analogous to that in Dia, which in turn was considered analogous to examining a photographic array;
(d)the identification occurred close in time to the offence;
(e)the complainant made a sworn statement about the identification process;
(f)there was no subsequent photoboard or digiboard identification and so the issue of displacement by further identification does not arise; and
(g)the complainant had ample opportunity to see the offenders during the offence and his vision was not compromised. It is not suggested his vision was affected by alcohol or the like and the outside lights were on (in contrast to concerns raised in Strauss v Police and Dia). Although it was dusk, the area was well lit. The complainant was able to see the offenders through glass doors.
I add that the State also says other circumstantial evidence to be taken into account includes a text or Facebook message passing from SA to Mr Roe prior to the offence and apparently referring to the complainant, in which she writes, 'Oh pip I seriously would like to organise something so he just leaves me the fuck alone once and for all. I wouldn't tell no body I give you my word ...'. The State also refers to evidence that E was apparently friends with Mr Roe.
I do not consider it appropriate to take this additional circumstantial evidence into account in determining the weaknesses or otherwise of the Facebook identification.

Determination

On its face, the identification evidence is probative. It goes directly to proving a fact in issue. While its weight may be limited, that is a matter for the jury.
Mr Roe's concerns as to prejudice fall generally within three categories: suggestibility, informality of process and lack of accuracy.
As to suggestibility, as made clear in Winmar, that is a matter that can be addressed by a direction. There is a degree of suggestibility in that the complainant started his search from a position that assumed the unidentified offender may appear on E's Facebook page, but I accept the State's submission that the risk of suggestibility is less than was the case in Peterson and Crawford and is similar to the risk addressed in Dia. In those cases the evidence was admitted and the risk associated with suggestibility was able to be dealt with by a direction.
As to the alleged lack of formality, for the reasons discussed in Peterson and contrary to comments in Strauss v Police, I do not consider a follow up photoboard or digiboard identification process was necessary: defence counsel also identified the risk of displacement if a further identification had been carried out and said the police acted quite wisely in not doing so.[35] I do not suggest that there may not be occasions where it is appropriate for the police to carry out a further identification process. I do not consider it was necessary in this case. I also note that the complainant explained the nature of his Facebook search to the police very shortly after it was undertaken. Both the complainant and the police officer have provided statements about the Facebook search and relevant photos and may be subject to cross examination as to that process at trial.
[35] ts 23.
As to accuracy, the full length photo of Mr Roe is in my view sufficiently clear to be able to say that there are resemblances between the person in that photo and the police photo of Mr Roe referred to by the defence. For example, there is a similarity in the shape of the face that can be seen even though in the Facebook photo Mr Roe is wearing a cap. Whilst ultimately a question for the jury, I do not agree with defence counsel that the photo from Facebook, when compared with the police photo, 'doesn't look like him at all'.[36]
[36] ts 20.
The fact that the complainant did not refer to any tattoos or other distinguishing features is not necessarily surprising: it is not known at present what the offenders were wearing on the day (except that one wore a bandana and the complainant says they both wore shoes) and whether any features such as tattoos would have been visible.
It is fair to say that the description in the complainant's statement of the person said to be Mr Roe is very general in nature. However, it is also fair to say that the complainant had a good opportunity, over a short period of time and apparently in good light, to observe the two offenders when they were both in the car and when they were immediately outside his house. For the reasons discussed in Winmar, the fact that the witness does not include a detailed description in his statement does not mean that the complainant is unable to recognise the offenders. Whether any direction is required on this point will be a matter for the trial judge, taking into account the evidence.
The issues that potentially arise on Facebook identification are not dissimilar to those that may arise with respect to other photographic means of identification. Dangers such as suggestibility, lack of formality, lack of opportunity or absence of a precise description are capable (where required) of being addressed by careful directions to the jury.
I consider the Facebook identification evidence in this case is probative and the risks associated with it can be dealt with properly by direction to the jury. The jury is able to assess its credibility and reliability. Accordingly, I would admit the evidence.