Identification parades:

1
Identification parades
2
Causes of miscarriages of justice  Borchard
(1932) 65 wrongful convictions. Most common
causes of error (a) mistaken identity (29, or
45, of convictions). In only two cases did the
defendant resemble the real culprit. (b)
over-reliance on circumstantial evidence. (c)
perjury by witnesses. (d) self-incriminating
confessions. (e) unreliability of expert
witnesses.
3
Brandon and Davies (1972) 70 British cases. (a)
Mistaken identification. (b) Self-incriminating
confessions.
4
Bedau and Radelet (1987) 350 cases in USA. Four
groups of errors - (a) Errors caused by police
investigation prior to trial (23 of sample). 49
(14) of these involved false confessions, mainly
due to coercion. (b) Errors caused by prosecution
prior to or during trial (50, or 14 of cases).
Most common error was suppression of evidence of
innocence. (c) Errors caused by prosecution
witness perjury (117 cases) and mistaken
identification (56 cases). (d) Miscellaneous -
circumstantial evidence (9), public demand and
outrage (20).
5
Eyewitness identification ID parades (lineups)
present the suspect amongst innocent foils
(distractors). Problem can be unreliable, due to
false positives (false identifications) and
false negatives (failing to identify the suspect
as present).
6

• Why are ID parades so unreliable?
• Unfamiliar face recognition is poor.
• Physical changes in the suspect between initial
  encounter and testing.
• 3. Witness misattribution of feelings of
  familiarity evoked by a suspects face.
• 4. Jurors have misplaced faith in witnesses
  confidence in their own identification accuracy.
• 5. Difficult to ensure that lineups are fair
  (unbiased).

7

• Neil vs. Biggers (1972)
• U.S. Courts five criteria for evaluating
  eyewitnesses
• Opportunity of the eyewitness to view the
  offender at the time of the crime.
• Witness degree of attention.
• Accuracy of the witness prior description of the
  offender.
• Level of witness certainty at the identification
  procedure.
• Length of time betwen crime and identification
  procedure.

8
Brown, Deffenbacher and Sturgill
(1977) Students saw 5 "criminals". 3 days
later, picked them from 15 mugshots. 7 days
later, picked them from 5 live line-ups. Identifi
cation rates in line-up phase 45 of "criminals"
seen both at crime and in mugshots . 18 of
individuals NOT seen at crime OR in mugshots
(false recognitions). 29 of individuals NOT seen
at crime, but seen in mugshots (false
recognitions). Exposure to mugshots increased
chances of false identification at line-up phase
familiarity but not context.
9
Wells, Lieppe and Ostrom (1979) Method for
assessing inter-item similarity within a lineup.
Non-witnesses try to identify suspect solely by
witness' verbal description. Should be at chance
line-up effective size is estimated from the
number of successful non-witnesses. Assessed the
line-up procedure in a real bank robbery case in
which a suspect was identified from a 6-person
line-up. By chance, 1/6 of non-witnesses should
be successful. In practice, 25/41 non-witnesses
identified the suspect. Effective size of the
line-up 41/25 1.64 - line-up distractors
could almost have been omitted!
10
Wogalter, Marwitz and Leonard (1992) Foils
selected on the basis of their similarity to the
suspect produce biased lineups (prototype
effects). Alternative methods are fairer (e.g.,
foils based on similarity to suspect and one
foil, or where all faces are equally similar to
each other).
11
Wogalter, Malpass and Burger (1993) Studied how
U.S. police officers actually produce
lineups. Similarity to suspect was main basis for
foil selection (which is bad). 94 decided
whether a lineup was "fair" on the basis of their
own judgement plus 77 got a second opinion from
another officer.
12
Simultaneous versus Sequential lineups Simultane
ous lineups allow witness to use a relative
judgement strategy i.e., to compare lineup
members, decide which is closest to their
memories for the criminal, and then infer this is
the guilty person. Result in higher rate of false
identifications, when the criminal is absent.
That's the one!
13
Simultaneous versus Sequential
lineups Sequential lineups force witnesses to
adopt an absolute judgement strategy - deciding
whether or not the person currently being
examined is the criminal. Sequential lineups
are superior to simultaneous lineups because they
reduce false-positive choices (Cutler and Penrod
1988 Lindsay and Wells 1985 Sporer 1993).
14
Steblay, Dysart, Fulero and LIndsay
(2001) Meta-analysis comparing sequential and
simultaneous lineups. Correct identifications
with target-present lineups simultaneous
50 sequential 35 Correct rejections with
target-absent lineups simultaneous
49 sequential 72
15
Accuracy and confidence Neil vs. Biggers
(1972) Witness confidence should be considered
in court. Little relationship between confidence
and identification accuracy (e.g. Luus and Wells
1994 Sporer, Penrod, Read and Cutler
1995). Bothwell, Deffenbacher and Brigham
(1987) confidence-accuracy correlations ranged
from 0.08 to 0.42. Wells and Murray (1984)
confidence/accuracy r 0.07.
16
Accuracy/confidence relationships and effects of
post-identification feedback Wells and
Bradfield (1998) Participants tried to identify
gunman from target-absent lineup. Group A
confirming feedback. Group B disconfirming
feedback. Group C no feedback. Feedback
distorted subsequent confidence ratings and
retrospective reports of how well they had seen
the gunman. Memon and Dixon (2005) Similar
study. Effects of negative feedback only
affected confidence but not accuracy.
17
Wells and Bradfield (1999) No post-identification
feedback effects if participants thought
privately about their certainty before obtaining
feedback. Luus and Wells (1994) "Perseverance
effects" of post-identification feedback.
Bradfield, Wells and Olson (2002) Confirming
feedback inflated retrospective certainty of
inaccurate witnesses but not accurate ones.
Bradfield et als results
Wells et al (2003) Accessibility hypothesis no
on-line evaluation of performance, so feedback is
best guide. Neuschatz et al (2005) Similar
effects for young and old participants supports
accessibility hypothesis rather than trace
strength explanation.
18
Douglass and Steblay (2006) Meta-analysis of
research on post-identification feedback 2,400
participants. Robust effects of confirming
feedback on retrospective certainty, quality of
view, memory and attention. Less powerful
effects of disconfirming feedback.
19

• Conclusions
• Eyewitness identification is highly influential
  with jurors, but a major cause of miscarriages of
  justice.
• Eyewitness confidence is highly influential with
  jurors, but is
• an unreliable guide to witnesses accuracy
• susceptible to alteration by post-identification
  feedback.

20

• Douglass and Steblay (2006)
• Recommendations for good lineups are
• Effective use of fillers.
• 2. Blind administration of lineup (PACE does not
  ensure this).
• 3. Warn witness that the culprit may or may not
  be present (reduces false identifications by 25
  - meta-analysis by Steblay 1997).
• 4. Sequential presentation.
• 5. Record eyewitness assessment of their
  certainty at the time identification is made.
• 6. Do not give witnesses feedback about their
  identification performance.
• Legal system should reconsider eyewitness
  evaluation procedures (e.g. advice to juries).
  Criteria of confidence, attentionand view
  should not be used if post-identification
  feedback was provided.