Atkins & Anor v R

A suitably qualified photographic-comparison expert may give an opinion on the significance of identified facial similarities or dissimilarities. The absence of a statistical database does not make that informed opinion inadmissible, provided it rests on demonstrated study and experience.

Hierarchical verbal expressions of support may be used, but should not be numbered or presented as a measurable scale. The jury must be told clearly that such expressions are subjective assessments, not positive identification or statistical proof. The evidence should be rigorously tested and accompanied by careful judicial directions.

The appellants, who were brothers, were convicted at the Central Criminal Court of murder, aggravated burglaries, wounding with intent and a firearms offence arising from two planned armed robberies. Each received life imprisonment with a specified minimum term of 35 years.

CCTV footage from the second robbery briefly showed one robber without a mask. A facial-comparison expert compared that image with photographs of the appellants and others. He excluded Michael Atkins and others, but could not positively identify Dean Atkins. He said that the similarities gave between support and strong support to the allegation that Dean Atkins was the man shown.

The appeals against conviction challenged only the admissibility of that expression of the expert’s conclusion. The applications concerning sentence challenged the length of the specified minimum term.

1. The appeals against conviction were dismissed. The applications for leave to appeal against sentence were refused.

2. The court’s common-law jurisdiction permits it, where admissibility is challenged, to require the party tendering expert opinion to establish both the relevant expertise and that the opinion is founded on it. The limited objection in this case did not challenge the expert’s qualifications or his comparison techniques.

3. A facial-comparison expert whose evidence is properly based on study and experience may express an opinion about the significance of similarities and dissimilarities. The absence of a statistical database does not confine the expert to describing the features and leave the jury with unsupported raw material. Such evidence cannot ordinarily make a positive identification, but it may support or exclude an alleged identification.

4. R v Gardner [2004] EWCA Crim 1639 was the directly relevant authority. The contrary observations in R v Gray [2003] EWCA Crim 1001 were obiter and had not been made after argument with the relevant authorities. The court accepted Gray’s caution that the lack of a database must be clearly explained, but rejected its suggestion that an expert must stop at identifying similarities.

5. Conventional ascending verbal labels, such as limited, moderate, strong or powerful support, may be used to communicate an expert’s subjective assessment. They should not be given numerical labels, which could imply a measurable scale. The jury must be directed that the assessment is subjective, is not statistically founded and is for the jury to evaluate. Disclosure, cross-examination and, where appropriate, contrary expert evidence provide important safeguards.

6. The sentencing starting point under Criminal Justice Act 2003, Schedule 21 was 30 years. Significant planning, a vulnerable victim and grave violence aggravated the murder. The judge was also entitled, applying totality, to reflect the separate violent aggravated burglary and wounding in a 35-year specified term. The term was neither wrong in principle nor manifestly excessive.

• Court of Appeal (Criminal Division) — The appeals against conviction were dismissed and the applications for leave to appeal against sentence were refused: [2009] EWCA Crim 1876.
• Central Criminal Court — His Honour Judge Kramer QC convicted both appellants of murder and the related offences, and imposed life sentences with specified minimum terms of 35 years.