RECENT RESEARCH

CODES OF PRACTICE FOR IDENTIFICATION

Wells, G.L., Small, M., Penrod, S., Malpass, R., Fulero, S.M. andBrimacombe, C.A.E. (1995) “Eyewitness Identification Procedures:Recommendations for Lineups and Photospreads”,Law and HumanBehavior22: 603–647.

DNA typing of human hair and tissue found at the scene ofthe crime has become a major weapon in the hands of the policeand the prosecution in unequivocally establishing the identity of theoffender. It has also become a powerful procedure in retrospectivelyproving the innocence of convicted persons who maintain they havebeen victims of miscarriages of justice. In this 1998 paper, 40 suchcases of proven wrongful conviction are described of which no lessthan 36 (or 90%) were based on identifications in which one ormore witnesses falsely identified the accused as the perpetrator ofthe crime. This sobering statistic led Janet Reno, the US AttorneyGeneral, to establish a panel, made up of lawyers, police officersand psychologists to draw up a set of guidelines for establishingidentification through line-ups (identification parades) or photo-spreads (identification from an array of mugshots). Photospreadsare by far the most common method used in the United States forestablishing identification, but are rarely, if ever, used in the UnitedKingdom. British witnesses may sometimes be shown mugshotalbums in a search for a suspect, but a positive identificationfrom an album is normally followed by a formal parade, thoughthere is a realisation that at the latter a witness may merely pickout the person whose mugshot they chose. The current paperrepresents an attempt by the American Psychology-Law Society(APLS) to provide a consensus statement on the current state ofknowledge on the psychological processes underlying eyewitness

Expert Evidence7: 59–65, 1999.© 2000Kluwer Academic Publishers. Printed in the Netherlands.

60 GRAHAM DAVIES AND TIM VALENTINE

identification and to provide a series of recommendations designedto minimise the possibility of wrongful convictions based on faultyidentification. The APLS working party led by Gary Wells, himselfa member of the Attorney General’s Panel, comprised many of theleading US researchers in the field: their opinions carry weight. It isinstructive therefore to consider their recommendations in the lightof relevant parts of the Police and Criminal Evidence Act (hereafterPACE) Code of Practice (Her Majesty’s Stationary Office, 1995) inuse in England and Wales.

The paper reviews the research literature, much of it published inthe last 20 years, which emphasises that identification from an array,whether of actual people or mugshots, is a relative rather than anabsolute process: the authors cite a study by Wells where witnessesselected the target correctly from an array on 54% of occasions.However, if in another condition, the target was removed from thearray, 68% of witnesses still selected a person from the array; themajority selecting the person most like the target in appearance. Theuse of instructions emphasising that the person ‘may or may notbe present’ serves to reduce but not to eliminate this effect. Theauthors go on to highlight that another effective way of reducingrelative choosing is to use a sequential rather than a simultaneousline-up. Under experimental conditions, forcing witnesses to makea judgement on each member of the parade in turn before seeing thenext member has no significant impact on correct identifications, butdrastically reduces the number of false positive responses.

The article reviews the vexed area of the relationship of eyewit-ness confidence and accuracy, noting in passing that American (andincidently, English) judges’ directions to juries recommend that theconfidence of witnesses in their identification should be a factor thatshould be taken into account. Sentiment among research psycho-logists has fluctuated over the years from the view that confidencewas not a reliable guide to accuracy to the view that high levels ofconfidence at the time of choosing may be a significant indicator ofactual accuracy. The authors note that the most exhaustive, recentmeta-analysis of previous studies found a modest but significantpositive correlation of 0.25. However, they go on to emphasise, cor-rectly in our view, that confidence can be manipulated independentlyof actual accuracy. The most dramatic demonstration of this is the

RECENT RESEARCH 61

impact of giving feedback to witnesses on their choice. Not onlydo witnesses who are told ‘you got him’ show greater retrospectiveconfidence in their identification, but equally alarmingly, they sub-sequently claim to have had a better view of the witness and to havepaid more attention to the suspect than do witnesses who had beentold they had incorrectly picked a stand-in (Wells and Bradfield,1998). All who have studied courtroom testimony will be familiarwith the inflation of confidence which seems to overtake witnessesfrom their initial statements to the police, through committal hear-ings, and thence to court, which appears to be independent of thereliability of the original testimony. If confidence is to be used bythe jury in reaching decisions, it needs to be the verbatim statementtaken at the time of the original identification, rather than what thewitness now says in the courtroom.

After such a well informed and full discussion of the psycholo-gical issues in identification, the actual changes to the rules whichthe paper recommends come as something of an anti-climax. Thismay be because of a self-imposed prohibition that any recom-mendations should not involve significant additional costs. The firstrecommendation is thatthe person conducting the parade shouldnot be aware of which member of the lineup or photospread is thesuspect. The authors’ concern is that an over-zealous officer who hasa commitment to solving the case may deliberately or accidentallybias the witness toward a particular choice. In England and Wales,the PACE Codes of Practice specify that a parade will be conductedby an officer of Inspector rank or above who is not a member ofthe investigating team. However, this ‘identification officer’ doesknow who the suspect is and in which position he is standing. Toensure an entirely double-blind procedure would require parades tobe conducted not by one but by two officers, one responsible forarranging the parade, the second for conducting the witness and not-ing down any choices and comments. Problems would still arise ifthere were multiple witnesses viewing the one parade. Under PACE,suspects have the right to change places between witnesses and toensure anonymity, responsibility for this re-arrangement would haveto be handed back to the second officer. Even so, the identifyingofficer, would still know who had been selected by the first witness.Logically, the only way of ensuring complete anonymity would

62 GRAHAM DAVIES AND TIM VALENTINE

be to allocate a different identification officer to each witness, anexpensive and time-consuming expedient. Police in the UK wouldno doubt argue that the provision in PACE for the presence of anaccused’s solicitor or representative provides sufficient protectionagainst the more obvious forms of malpractice.

The second recommendation is thateyewitnesses should be toldexplicitly that the person in question might not be in the lineup orphotospread and therefore should not feel that they must make anidentification. They should also be told that the person administer-ing the lineup does not know which person is the suspect in the case.The PACE instructions anticipate this recommendation: “Immedi-ately before the witness inspects the parade, the identifying officershall tell him that the person he saw may or may not be on the paradeand if he cannot make a positive identification he should say so, butthat he should not make a decision before looking at each mem-ber of the parade at least twice”. (p. 95, PACE Codes of Practice).There is certainly evidence that witnesses do take advantage of thisprovision when it is offered. Wright and McDaid (1996) reportingon a sample of 1561 parades held in London in 1992, found that39% of witnesses picked the suspect, 20% picked a foil and no lessthan 41% did not make a selection. The problems of ensuring thatthe identification officer does not know who is the suspect are dealtwith above.

Recommendation 3 deals with the relative appearance of sus-pect and foils:the suspect should not stand out in the lineup orphotospread as being different from the distracters based on eyewit-ness’ previous description of the culprit or based on other factorswhich would draw attention to the suspect. This emphasis uponbuilding a parade around persons who resemble the verbaldescrip-tion of the suspect contrasts with the English tradition, reflected inPACE, for the foils to resemble the suspect’s physicalappearance“in age, height, general appearance and position in life” (p. 100,PACE Codes of Practice). The same code, however, does demandthat the witness’s initial description of the suspect be provided inwriting to the suspect or his/her legal representative. Given that alegal representative is present, it would always be possible for himor her to object to any stand-ins who departed significantly fromthe verbal description in much the same way as currently the legal

RECENT RESEARCH 63

representative can object to any person who is markedly different inphysical appearance from the accused. The new emphasis placed onresemblance to verbal descriptions does have practical drawbackscompared to the traditional method. First, verbal descriptions canvary from the extremely precise through to the most vague andgeneral: a reliance purely on description in these contrasting cir-cumstances could produce parades with wide differences in degreeof similarity between individual foils and between foils and thesuspect. Second, where multiple witnesses were involved, separateindividual parades would have to be constructed correspondingto the description of each witness, rather than the same paradebeing used for all witnesses, as at present. Here, again, this recom-mendation may reflect the North American practice of favouringphotospeads over live lineups for identification purposes.

The final recommendation that “a clear statement should betaken from the eyewitness at the time of the identification and priorto any feedback as to his or her confidence that the identified per-son is the identified culprit” reflects recent research on the valueof contemporary, as opposed to retrospective judgements of confid-ence. This is an area of ambiguity in the current PACE Codes whichdeserves to follow the recommended North American practice. Thecurrent parade documentation to be completed by every identifica-tion officer in England and Wales includes provision for taking downthe verbatim remarks of the witness and these can include expres-sions of confidence. It is also common practice where witnessesmake a positive identification at a parade, for them to make a furtherstatement to the police if they showed any hesitation and expressingtheir confidence in their judgement. Given that such statements aremade after feedback on choice, their evidential value is problematicand a formalisation along the lines proposed by the APLS WorkingParty would be a useful addition to PACE.

The final section of the paper is devoted to two proceduralimprovements which were rejected as “too many rules would pro-duce resistance from police and legal policy makers” (p. 639). Thefirst is for the introduction of sequential parades. After detailingtheir many advantages over conventional parades where all mem-bers are simultaneously present, the authors decide not to makea recommendation in part because sequential parades rely “on

64 GRAHAM DAVIES AND TIM VALENTINE

more complex understanding of the problem based on the relative-judgement conceptualisation that we do not think is a part of theintuitions of legal policy makers at this point” (p. 640). We believethis may underestimate not only the conceptual capacities of USlaw enforcement personnel, but also the communicative abilities ofpsychologists interested in applying psychology in a legal context.Procedures in England and Wales currently do not favour sequentialparades. However, many forces are switching from live to videoparades, at which witnesses view video clips of volunteer stand-ins,together with a similar clip of the accused. Currently, witnesses viewall the clips twice before making a decision. Given that the clips arepresented sequentially, it would be relatively straightforward changeto ask the witness for a response after each clip.

A final rejected recommendation concerns the videotaping ofparades, which as the authors note, would provide a permanentrecord not only of the structure of the parade, but also of what wassaid and by whom. This proposal too, is rejected on the basis of costand practicality. This rejection elicits the only departure from unan-imity with Kassin arguing strongly for its inclusion in a minorityreport (Kassin, 1998). In support of this proposition, Kassin statesincorrectly that PACE requires all interviews with suspects to bevideotaped (this should of course refer to audiotaping). He could,however, have pointed to the many examples of customised iden-tification suites in London and elsewhere in the United Kingdomwhere parades are routinely videotaped. This represents a markedimprovement on the minimum requirements of PACE which spe-cifies that an identification parade should always be photographedin colour.

In conclusion, with the exception of the recommendation onadopting a double-blind procedure and formally recording confid-ence, existing regulations on identification in England and Waleseither anticipate these recommendations or provide safeguardswhich go beyond their very modest proposals. In particular, NorthAmerican practice has something to learn from UK practice of per-mitting legal representation at parades and video identifications andin the increasing practice of videotaping parades. The latter hasprobably spread so rapidly as it is perceived as having advantagesfor both prosecution and defence: the defence have a record of

RECENT RESEARCH 65

what went on, while the police can use it to show that the paradewas fair and procedures followed to the letter. However, the appar-ent superior position under PACE guidelines should not encouragecomplacency. Mistaken identification continues to be a significantsource of miscarriages of justice in England and Wales (Davies,1996). However, such errors are now more likely to reflect a fail-ure to follow the existing Code of Practice, rather than to gaps orambiguities in the Code itself.

REFERENCES

Davies, G.M. (1996) “Mistaken Identification: Where Law Meets PsychologyHead on”,The Howard Journal35: 232–241.

Her Majesty’s Stationary Office (1995)Police and Criminal Evidence Act 1984(s.60(1)(a) and s.66) Codes of Practice. London: HMSO Publications.

Kassin, S.M. (1998) “Eyewitness Identification Procedures: The Fifth Rule”,Lawand Human Behavior22: 649–653.

Wells, G.L. and Bradfield, A.L. (1998) “Good, You Identified the Suspect: Feed-back to Eyewitnesses Distorts their Reports of the Witnessing Experience”,Journal of Applied Psychology83: 360–376.

Wright, D.B. and McDaid, A.T. (1996) “Comparing System and Estimator Vari-ables Using Data from Real Line-ups”,Applied Cognitive Psychology10:75–84.

GRAHAM DAVIES TIM VALENTINE

Department of Psychology Department of PsychologyLeicester University Goldsmiths CollegeUK London University

UK