Handbook of Psychology of Investigative Interviewing: Current Developments and Future Directions
Page 40
14 , 326 – 329 .
Wegner , D. M. , Schneider , D. J. , Knutson , B. & McMahon , S. R. ( 1991 ). Polluting
the stream of consciousness: The effect of thought suppression on the mind ’ s
environment . Cognitive Therapy and Research , 15 , 141 – 152 .
Weingardt , K. R. , Loftus , E. F. & Lindsay , D. S. ( 1995 ). Misinformation revisited:
New evidence on the suggestibility of memory
.
Memory and Cognition ,
23 ,
72 – 82 .
Wenar ,
C. (
1961 ).
The reliability of mothers
’ histories
.
Child Development ,
32 ,
491 – 500 .
Wenar , C. & Coulter , J. B. ( 1962 ). A reliability study of developmental histories . Child
Development , 33 , 453 – 462 .
Wilkinson , C. & Hyman , I. E. ( 1998 ). Individual differences related to two types if
memory errors: Word lists may not generalize to autobiographical memory
.
Applied Cognitive Psychology , 12 , S 29 – S 46 .
Wilson , K. & French , C. C. ( 2006 ). The relationship between susceptibility to false
memories, dissociativity, and paranormal belief and experience . Personality and
Individual Differences , 41 , 1493 – 1502 .
Winograd , E. , Peluso , J. P. & Glover , T. A. ( 1998 ). Individual differences in suscep-
tibility to memory illusions . Applied Cognitive Psychology , 12 , S 5 – S 27 .
Wright , D. B. , Ost , J. & French , C. C. ( 2006 ). Ten years after: What we know now
that we didn ’ t know then about recovered and false memories . The Psychologist ,
19 , 352 – 355 .
Wrightsman , L. S. & Kassin , S. M. ( 1993 ). Confessions in the courtroom . London : Sage .
Zaragoza , M. S. , McCloskey , M. & Jamis , M. ( 1987 ). Misleading postevent informa-
tion and recall of the original event: Further evidence against the memory impair-
ment hypothesis
.
Journal of Experimental Psychology: Learning, Memory, and
Cognition , 13 , 36 – 44 .
Chapter Twelve
Obtaining and Interpreting Eyewitness
Identifi cation Test Evidence: The
Infl uence of Police – Witness Interactions
Neil Brewer
School of Psychology
Flinders University
and
Gary L. Wells
Department of Psychology
Iowa State University
Introduction
Eyewitnesses to a crime are frequently asked to view an identifi cation parade
to see if they can identify the offender. Conduct of a line - up involves police
or line - up administrators in a number of important decisions, such as who to
put in the line - up, the method of presentation of the line - up, and what to say
to witnesses before and after the line - up. The identifi cation test can be con-
ceptualized as a variant on an interview between the police and the witness,
involving important interactions between police (or other line - up administra-
tors) and witnesses. These interactions can profoundly infl uence witness deci-
sions and impact on the characteristics of any subsequent evidence they provide
in the courts. We shall focus on (i) the expectations that police/administrators
can engender in witnesses and how these can shape witness behaviour; (ii) the
instructions that are provided to witnesses prior to viewing the line
- up;
(iii) possible ways in which administrators can interact with (and hence infl u-
ence) witnesses in the conduct of line - ups; (iv) the soliciting of confi dence
Handbook of Psychology of Investigative Interviewing: Current Developments and Future Directions
Edited by Ray Bull, Tim Valentine and Tom Williamson
© 2009 John Wiley & Sons, Ltd.
206
Handbook of Psychology of Investigative Interviewing
assessments from witnesses; (v) the interpretation of witness confi dence assess-
ments; and (vi) the infl uence of interactions that occur post - identifi cation on
witnesses ’ subsequent reports about the event and the identifi cation test.
Eyewitnesses to crimes are frequently asked by police to view an identifi ca-
tion parade or line - up to see if they can identify the offender who, of course,
may or may not be the police suspect. It is now well documented that eyewit-
nesses are far from infallible when it comes to making an identifi cation, with
their performance characterized by both mistaken identifi cations of innocent
suspects and failures to identify the offender when present in the line - up (e.g.
Cutler & Penrod, 1995 ; Wells, Memon & Penrod, 2006 ; Innocence Project,
2009 ). In sum, eyewitness identifi cation tests are far from a fool - proof way of
pinpointing the offender.
It is also well established that the conduct of a line - up requires police or
other line - up administrators to make a number of decisions that can have a
far - reaching impact on the outcomes of the identifi cation test. These decisions
include deciding upon the membership of the line
- up (e.g., how many
members, how closely those members resemble the appearance or the descrip-
tion of the perpetrator), the method of presentation of the line - up (e.g., photo
spread vs. live line - up vs. video line - up; simultaneous vs. sequential), how to
record the line - up decision and any associated behaviours of the witness (e.g.,
witness confi dence, decision latency), what to say to witnesses before they are
shown the line - up and after they have made their decision (e.g., feedback
about their decision). The research literature on these issues is extensive and
offers valuable guidelines for the conduct of line - ups (e.g. Technical Working
Group for Eyewitness Evidence, 1999 ; Brewer, Weber & Semmler, 2005 ;
Wells et al. , 2006 ), but many of these issues are not the focus of this chapter.
In this chapter we conceptualize the identifi cation test as a special form of
interview between the police and the witness. We examine the important
interactions that can occur between police (or other line - up administrators)
and witnesses. We show how these interactions can exert a signifi cant infl uence
on witness decisions and behaviour and, in turn, shape the characteristics of
any subsequent evidence they provide (e.g., in the courts).
Pre - i dentifi cation t est i nteractions
An identifi cation test is seldom likely to take place immediately after a crime.
Rather, a more likely scenario is that police will fi rst identify witnesses to the
crime and interview them about the participants and events associated with the
crime. At a later date (in the UK after a delay likely to exceed one month; Pike,
Brace & Kynan, 2002 ) when police have a suspect, they may ask a witness to
attend an identifi cation test. There is some empirical evidence, and there are
sound theoretical grounds, for believing that interactions that occur prior to
the identifi cation test may shape the responding of the witness at the identifi ca-
tion test.
Obtaining and Interpreting Eyewitness Identifi cation Test Evidence 207
Effects of p roviding a d escription on
i dentifi cation p erformance
It is possible that providing a verbal description to police at an interview may,
under certain circumstances, impair subsequent performance on an identifi ca-
tion test, an effect labelled the verbal overshadowing effect (Schooler
&
Engstler - Schooler,
1990 ). Although Meissner
& Brigham
’ s
(2001) meta
-
analysis of verbal overshadowing studies revealed a small negative effect of
providing a description on recognition, the fi ndings are certainly not univer-
sally supportive. Further, while the effect has been detected in experiments in
which the identifi cation test closely followed the verbal description (e.g., ≤ 10
minutes) – conditions that are unlikely to prevail in real investigations – a mild
facilitation effect has been detected when this interval was extended (ibid.). It
is probably a fair assessment of the state of the scientifi c literature in this area
to say that the fi ndings are not decisive, the effect sizes appear not to be large
and the key theoretical mechanisms underpinning the effects are by no means
resolved. From a practical perspective, the possible existence of the verbal
overshadowing effect can perhaps be ignored, as the occurrence of a police
interview with a witness early in an investigation, and prior to any identifi ca-
tion test, would seem to be inevitable. Should, however, future research reveal
a facilitation effect to be robust under conditions paralleling those likely to be
found in real investigations, it is possible that clarifying the mechanisms under-
lying such effects could assist in the refi nement of interviewing techniques that
might facilitate identifi cation performance.
Another possible effect of a prior interview on identifi cation performance
has been described by Brewer (2006) . It is possible that the performance of a
witness during a police interview will shape the witness ’ s impressions of the
quality of his or her memory, with these impressions affecting the likelihood
that she or he will make a choice at the identifi cation test. For example, it
seems intuitive that a witness who had great diffi culty describing the offender
at interview may doubt the quality of their memory and be less likely to choose
from the line - up; the opposite might be expected for a witness for whom the
description seemed particularly easy. Witnesses ’ inferences about the strength
of their memory for the perpetrator might also be shaped by any post - interview
feedback provided by police interviewers.
Interestingly, however, the basic recognition memory literature suggests
the possibility of a different, and counter - intuitive, effect. Word - recognition
research conducted within a signal detection framework indicates that people
who are likely to believe that they have a strong memory for particular stimuli
(because they have studied the stimulus items extensively) expect to be able
to remember those stimuli and set a more demanding criterion for reporting
their occurrence (Stretch & Wixted, 1998 ; Morrell, Gaitan & Wixted, 2002 ).
Conversely, people who believe the opposite (i.e., a weak memory) relax their
criterion and are more likely to report that they have seen a stimulus before
208
Handbook of Psychology of Investigative Interviewing
(Stretch & Wixted, 1998 ). Clear evidence for the infl uence of such witness
metacognitions has not yet been published in the scientifi c literature. However,
preliminary results from research currently underway in our laboratory reveal
patterns consistent with the latter, counter - intuitive hypothesis.
Expectations of w itnesses a bout the i dentifi cation t est
Most witnesses are likely to have limited experience with, or knowledge of,
identifi cation tests. Indeed their background probably derives mainly from
viewing television programmes involving police and lawyers. Nevertheless,
what witnesses almost certainly appreciate is that somewhere in the line - up is
a police suspect. Thus, when witnesses are contacted and asked to attend an
identifi cation test, they are probably going to infer that the police investigation
has led them to pinpoint a suspect. To the extent that a witness believes that
the suspect is likely to be the offender, the witness may also reason that a sign
of a capable witness will be the ability to pick the suspect from the line - up,
thereby assisting in bringing that person to justice.
While we are not aware of any data that provide an unambiguous indication
of witness beliefs on this issue, Memon, Gabbert & Hope (2004) reported
that more than 90% of participant witnesses across multiple experiments indi-
cated that they had assumed the perpetrator ’ s presence in the line - up, despite
having received line - up instructions explicitly warning them that the perpetra-
tor may not be present. These estimates involved retrospective reports from
witnesses who had just seen a line - up and may not, therefore, refl ect witnesses ’
a priori expectations about the presence of the perpetrator. Unpublished data
from our laboratory indicate that approximately 50% of university students
(sampled when attending the laboratory to participate in other non - eyewitness
studies) believe that more than 70% of ‘ real ’ police line - ups are likely to contain
the perpetrator. Around 75% believe that at least 50% of line - ups are likely to
contain the perpetrator.
Despite the limitations of such surveys, they reinforce the position that wit-
nesses who attend an identifi cation test are likely to entertain the belief that
the perpetrator is in the line - up, a belief that is likely to bias the witness towards
making a choice from the line - up. This bias can readily be enhanced if the
police or line - up administrator provides, knowingly or unknowingly, addi-
tional cues to the witness. One example would be if the witness was contacted
and told, ‘ We know who did it, but we need you to come view a line - up. ’
But, it does not have to be this blatant. For instance, the police might say,
‘ We have made great progress on the case and now we need to show you a
line - up ’ . Even the phrase ‘ We would like you to come to the station and see
if you can identify the perpetrator ’ implies to the witness that the perpetrator
is in the line - up and that the only question is whether the witness is capable
of picking him out. Interestingly, raising the expectation that the perpetrator
will be in a later line - up need not come from a conversation about a line - up
at all. For example, ‘ This guy has done this type of offence before and we need
Obtaining and Interpreting Eyewitness Identifi cation Test Evidence 209
to get him off the streets ’ is the type of statement that communicates to the
listener that the police know who did it. Hence, if a line - up is shown later,
the witness is going to assume that the police have the villain in the line - up;
otherwise, why are they showing the line - up?
Given the overwhelming evidence from laboratory and fi eld studies of wit-
nesses ’ propensity to misidentify innocent suspects, ensuring that witnesses
about to attend a l
ine - up are not biased towards making a positive identifi ca-
tion becomes a crucial part of the interaction between the line - up administra-
tor and the witness.
Police – w itness i nteractions at the i dentifi cation t est:
p re - d ecision i nfl uences
During the administration of the identifi cation test, abundant opportuni-
ties exist for the line
- up administrator to infl uence the decision - making of
the witness and the judicial outcome of the test. Some of these opportunities
occur prior to the witness actually making their decision about the line - up;
others occur after the decision. Consequently, eyewitness researchers insist that
certain procedures should be adopted to protect the witness from being infl u-
enced by the line - up administrator. At the pre - decision stage there are two
important procedures to follow: one involves what is called double
- blind
line - up administration; the other involves the use of unbiased line
- up
instructions.
Double-blind administration
Double - blind administration means that the line
- up administrator has no
knowledge of which line - up member is suspected of being the perpetrator and
which line - up members are merely fi llers. Fillers are known - innocent members
of the line - up who fi t the same general description but clearly are not the
perpetrator. The purpose of the fi llers is to prevent the witness from knowing
which person the police suspect and to make the witness rely on his or her
memory instead. Obviously, if the line - up administrator somehow communi-
cates to the eyewitness which line - up member is the suspect (and/or which
are mere fi llers), then the entire idea of using fi llers is undermined. The use
of a double - blind line - up administrator to prevent such communication for
eyewitness line - ups was fi rst proposed more than 20 years ago (Wells, 1988 ),
but the idea of double - blind test administration is a long - established staple for
human testing protocols in basic and applied science. The pharmaceutical
industry, for example, is required to use double - blind testing for new drugs
in which the medical testers of the patients cannot know whether the patient
is in the placebo condition or the drug condition because the testers might