The Court of Appeal gave judgment on the 18th December in the case of Donovan & Kafunda  EWCA Crim 2749. Here they gave their reasons for quashing the convictions for Mr Donovan and Mr Kafunda for a robbery and violent disorder committed during last year’s riots. The facts of the robbery were well documented in the media at the time.
Facts and Decision
There was CCTV and mobile phone footage of the incident, but the faces of the perpetrators could not be seen and the Prosecution accepted that it was not a case where any scientific process (for example, facial mapping) could assist.
The only other evidence was ‘cell-site’ evidence – evidence that mobile phones used by the defendants were in the general area where the offence was committed (as were many other people) and that the defendants had lied to the police about not being in that area when they were interviewed by the police (perhaps understandably, and in any event, this is not positive evidence that they were guilty). It was accepted that none of this would be anywhere near enough to get a conviction.
So what else was there that meant there could be a trial? The case against both men turned on the identification of one (different) witness each. Both of these witnesses stated that they were too scared for their identities to be known and were granted Witness Anonymity Orders. The provision for these orders were (re)introduced by s86 Coroners and Justice Act 2009 after the House of Lords ruled that there was no power for the identity of a witness to be withheld from the defence. This meant that in addition to the defendants not knowing the identity of the witnesses, the witnesses were screened and their voices disguised so that they could not be identified.
The witness who implicated Mr Donovan was called ‘Kieran Thomas’ (a pseudonym) and ‘Sarah Bishop’ identified Mr Kafunda (another pseudonym). Seemingly, they were not present at the attack, but were able to identify the two defendants from the CCTV footage based of their long knowledge of them.
Kieran Thomas was sure of his ID based on :
- Clothing (a pair of trainers with an ‘unusual’ green stripe, cap and Ralph Lauren jacket)
- The ‘way he smoked a cigarette’
Sarah Bishop was sure of her ID because of :
- Recognition of Mr Kafunda’s profile
- The way he was dressed
- The ‘way he helped [the victim] to his feet was characteristic of’ Mr Kafunda
The Court of Appeal noted that neither of these identifications “provide[d] overwhelming or compelling evidence of recognition or identification” and was “on any view … not of the best”. The weakness of the identification did not of itself feature as a ground of appeal however.
The issue for the Court of Appeal was whether the Witness Anonymity Orders were validly made. These should not be made routinely (see Mayers  EWCA Crim 2989). The reason for this is that it is on the face of it manifestly unfair for someone to be convicted on the basis of evidence from a witness who’s identity is unknown.
The actual decision of the Court of Appeal seems to have been on narrow grounds relating to a short passage in the (redacted) witness statements of each eyewitness that seemed to give them both a motive for lying about the defendants. This would have been before the Judge, but it is not clear whether or not his attention was specifically directed to it. Given the exceptional nature of Witness Anonymity Orders, it seems that the Court of Appeal gave the case more scrutiny (and the trial Judge less leeway and deference) than would be normal. Having reviewed everything, the Court concluded that the Orders shouldn’t have been made. In light of that, the Appeal was allowed.
The result of the Court of Appeal is clearly correct. Whilst the comments as to what was in the statement that triggered the suspicions aren’t clear, it is clear that there were substantial credibility issues with the two witnesses. It is pleasantly surprising to see the Court being as willing as it was to go behind the decision of the trial Judge.
What maybe more surprising to the general public is that even without those credibility issues there could be a conviction at all. Is it really possible to identify someone from the way they smoke a cigarette or help another person to their feet? It does seem, frankly, unlikely.
Identification is a notoriously complicated area with the potential for miscarriages of justice, even with honest witnesses trying to do their best. Additionally, people often vastly over-estimate their ability to recognise others. Whilst this has been long recognised, there has been a great deal of research been done on this point, in particular over the last 40 years. Regrettably, this has not often been taken on board by the Court. This will form the basis of a separate pRegrettably, this has not often been taken on board by the Court. This will form the basis of a separate piece, but there is a very good analysis of this from America last year.
It was not necessary for the Court of Appeal to consider the problems with the identification in any detail, but it would have been possible to have figures for how common the clothing is and, more importantly, the results of experiments to show just how possible it is to identify gait, profile and other matters. The sense from the judgment (although of course no transcript of the trial or other papers are available) is that this was a somewhat optimistic prosecution. In other cases, however, this may be much more of a live issue.
Science, and more particularly the scientific method, works. And for this reason, it has infiltrated all areas of our lives. Maybe it’s time that the Courts took a bit more notice?