laboratory that have addressed eyewitness memory and truthfulness/decep-
tion, with particular attention being given to the paper presented by myself
and Elizabeth Wells and Linda Yuval at the 2nd International Investigative
Interviewing Conference held at the University of Portsmouth, July 2006.
Attention will be given to the following issues:
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.
286
Handbook of Psychology of Investigative Interviewing
1. Based on common
- sense understandings of eyewitness behaviour, how
would men and women anticipate their likely responses as truthful or
deceptive eyewitnesses when questioned regarding a violent crime?
2. What is the relationship between truthfulness/deception and eyewitness
identifi cation?
3. Can university students distinguish between deceptive and truthful eyewit-
nesses in their testimony?
4. Are there differences between deceptive (research) suspects and honest
suspects in their repeated recall of real - world personal activities, including
their participation in either a crime or non - crime incident, respectively?
5. Can duration estimations be used as an indicator of deception?
Common s ense and d eception in e yewitness r eports
The objectives of police interviews are to determine whether a crime has
occurred, to obtain accurate descriptions and evidence that can identify the
individual(s) involved and to determine the credibility and truthfulness of the
eyewitnesses and suspects (Gudjonsson, 1992 ; Kebbell & Wagstaff, 1997 ).
Most deceptive witnesses, of course, will attempt to be persuasive and credible
in their reports to the police. In contrast to the reports of truth - tellers, the
reports of deceptive witnesses will probably be based in part on logical or
common - sense understandings of eyewitness memory and on their beliefs
about how they could best protect themselves from being found out. Previous
studies of common knowledge of eyewitness memory have followed various
approaches, including: the use of questionnaires (e.g., Kassin, Tubb, Hosch
& Memon,
2001 ; Benton, Ross, Bradshaw, Thomas
& Bradshaw,
2006 );
written descriptions of trials or videotaped trials (e.g., Hosch, Beck
&
McIntyre, 1980 ; Wells, Lindsay & Tousignant, 1980 ); prediction studies of
the accuracy of eyewitness identifi cation in staged crimes (e.g., Leippe, Wells
& Ostrom, 1978 ; Kassin, 1979 ); and the cross - examination of eyewitnesses to
staged crimes (e.g., Wells, Lindsay
& Ferguson,
1979 ; Lindsay, Wells
&
Rumpel, 1981 ). These different procedures have found relatively consistent
results, that is, laypersons (potential jurors), as well as police offi cers, lawyers
and trial judges, have poor understanding of many issues involved in eyewit-
ness memory.
Unlike the above studies, Yarmey (2004) asked participants how accurate/
inaccurate they believe they would be as honest or deceptive eyewitnesses,
respectively, in a criminal situation. Some 219 men and women in three
Canadian cities were asked how they would respond if coerced to lie to protect
a male perpetrator and female perpetrator, in contrast to being truthful and
as accurate as possible. Participants were given a description of an armed
robbery involving the shooting of a store worker, including a description of
the persons involved. The incident was described as lasting either 15 seconds
Truthfulness in Witnesses’ and Suspects’ Reports
287
or two minutes, and participants were told to assume they were the sole eye-
witness to the crime. Participants assigned to the deceptive condition were
told to assume that they and their family would be harmed if they did not
protect the perpetrator in their eyewitness reports. Participants assigned to the
truthful condition were asked to respond as they would when trying to be as
accurate as possible. Truthful and deceptive participants were instructed to
state their expected accuracy/distortion of reports in estimating the height,
weight and age of the two perpetrators, their estimation of the duration of the
incident and their identifi cation decision regarding a male suspect in a photo
line - up.
The results indicated that both truthful and deceptive witnesses had a bias
towards overestimation rather than underestimation of the height, weight and
age of the two perpetrators. As expected, the bias was greater for deceptive
witnesses than for truthful witnesses. Note, however, one interesting sex
difference: the data indicated that if coercive threats were made to self and
family, women in the deceptive condition would be more likely than their male
counterparts to signifi
cantly distort their reports regarding perpetrator
characteristics.
Prior to making their identifi cation decisions, all participants were told to
assume that the photo of the perpetrator was present in the photo line - up.
Most truthful witnesses expected to be highly accurate (84% hits, 2% false
alarms, 1% not there, 13% don ’ t know) in their photo identifi cation of the
male suspect. In contrast, most deceptive witnesses indicated that they would
say either that the suspect was not present in the line - up (31%), or that they
did not know if he was present (37%), rather than falsely identify an innocent
foil (4%). Truthful witnesses also expected to be signifi cantly more confi dent
in their identifi cation decisions than deceptive witnesses. Contrary to expecta-
tions, a substantial minority (27%) of deceptive witnesses said they would
correctly select the perpetrator in spite of possible retribution to both them-
selves and their family. Deceptive witnesses, in contrast to truthful witnesses,
reported signifi cantly less confi dence in their identifi cation decisions. The reac-
tion time to make identifi cation decisions also revealed signifi cant differences
between truthful and deceptive witnesses. Truthful witnesses indicated that
their identifi cation decisions would be signifi cantly faster than did deceptive
witnesses.
Interestingly, participants on average in both the truthful condition and the
deceptive condition indicated that they would overestimate the 15
- second
criminal event by 306%, and the two - minute incident by 88%, with no signifi -
cant differences between truthful and deceptive participants. Of interest is the
fact that several experiments show that short - duration events are signifi cantly
overestimated, and that women make signifi cantly larger overestimations than
do men (e.g., Loftus, Schooler, Boone & Kline, 1987 ; Yarmey & Yarmey,
1997 ). Consistent with the research literature, this study showed that women
indicated that they would overestimate the duration intervals at better than a
2 : 1 ratio than would men. It appears that women more than men may both
288
Handbook of Psychology of Investigative
Interviewing
anticipate and then act in crime - like situations with the expectation that time
will appear to move very slowly when they are witnesses to a highly emotional
criminal event.
Truthfulness and d eception in e yewitness i dentifi cation
In a laboratory - designed experiment, Parliament & Yarmey (2002) presented
undergraduate psychology students with a videotaped staged crime of a young
man abducting an eight - year - old boy. Participants were randomly assigned
to a deception group or a truthful group. Deceptive participants were instru-
cted to lie to protect the perpetrator, whereas truthful participants were told
to tell the truth and to identify the perpetrator to the best of their ability.
One week later all participants were given either a seven - person sequentially
presented suspect - absent photo line - up or suspect - present photo line - up. Of
the 16 deceptive participants given the suspect - present line - up, 100% claimed
the perpetrator was not there. In contrast, in the truthful group, 11 of 16
participants (69%) correctly identifi ed the target, four identifi ed foils and one
said he was not present. Given a suspect - absent line - up, 14 of 16 deceptive
participants (88%) said he was not present and two made false identifi cations.
In contrast, seven of 16 truthful witnesses (44%) correctly rejected the line - up
and nine (56%) made false identifi cations. Deceptive participants also made
their identifi cation decisions signifi cantly faster than truthful participants in
both the suspect - present and the suspect - absent line - ups.
In summary, these results suggest that deceptive eyewitnesses will behave
signifi cantly differently from truthful witnesses in their identifi cation decisions
and reaction times to make decisions. These results are interesting insofar as
they suggest that eyewitness identifi cations are not consistent with how indi-
viduals believe they would behave in a traumatic criminal situation with respect
to their identifi cation decisions as either honest or deceptive eyewitnesses, as
demonstrated in the questionnaire study on common sense just described.
Future research is needed to determine the limits of these differences in a
variety of situations.
Judgements of d eception and a ccuracy of
p erformance in e yewitness t estimony
Can university students distinguish between eyewitnesses who were deceptive
or truthful in their testimony, and either accurate or mistaken in their iden-
tifi cation of a target in a photo line
- up (Desmarais
& Yarmey,
2004 )?
Participants, all of whom could be potential jurors, were shown videotapes
taken from an unpublished study (Yarmey,
2003 ) of four truthful and
four deceptive eyewitnesses as they were being interviewed about their descrip-
Truthfulness in Witnesses’ and Suspects’ Reports
289
tions and photo identifi cation of a perpetrator in a fi eld experiment. Two
witnesses in each group were accurate and two were inaccurate in their line -
up identifi cation as determined by post - experimental analysis. (Note that two
of the deceptive eyewitnesses were correct in their identifi cation of the per-
petrator in spite of their assigned role to ‘ protect ’ the suspect. This may have
occurred because they forgot who the suspect was and mistakenly identifi ed
him, or refused to maintain their role as deceptive eyewitnesses and correctly
identifi ed the suspect.) Students rated each of the eight videotaped eyewit-
nesses on his or her accuracy of testimony, honesty, competence, confi dence
and credibility, and also indicated which verbal and nonverbal cues guided
their ratings.
Participants ’ decisions regarding deception and accuracy of eyewitness iden-
tifi cation were not signifi cantly different from what would be expected by
chance. However, participants were signifi cantly better in detecting honesty in
truthful eyewitnesses than they were in detecting lying by deceptive eyewit-
nesses. Also, participants were signifi cantly better in detecting accurate eyewit-
nesses than inaccurate eyewitnesses.
In terms of subjective impressions, truthful eyewitnesses were perceived as
more honest and more credible than deceptive eyewitnesses. Also, truthful
eyewitnesses were perceived to be more honest and more credible when they
were correct in their identifi cation. In contrast, deceptive eyewitnesses were
considered more honest and more credible when they were incorrect in their
identifi cations. Thus, deceptive eyewitnesses appear to be able to project cred-
ibility and trust when they misidentify a target.
Participants reported that they relied on different verbal and nonverbal
cues when judging truthful and deceptive eyewitnesses, and accurate and
inaccurate testimony. For example, participants relied on pauses or hesita-
tions in speech, frequent head movements and postural shifts to charac-
terize deceptive witnesses. In contrast, truthful witnesses reportedly were
discriminated by faster rates of speech, greater amounts of recalled details
and higher frequency of smiles and laughter. Different cues also were used
to differentiate between accurate and inaccurate eyewitnesses. Witnesses
were judged more likely to be accurate if they spoke quickly, gave many
descriptions, gave descriptions that were judged relevant and showed frequent
arm and hand movements. In contrast, witnesses were judged to be inaccurate
if they showed frequent pauses or hesitations in speech, admitted their forget-
fulness or inadequacy and displayed frequent smiles and laughter. Eye contact
or gaze were not said to be relied on to differentiate between truthful and
deceptive eyewitnesses, whereas some studies have shown that liars engage in
more eye contact than do truth - tellers (e.g., Vrij & Easton, 2002 ; Str ö mwall
& Granhag, 2003 ).
It may be concluded that people use different verbal and nonverbal cues to
differentiate between truthful and deceptive eyewitnesses, and between accu-
rate and inaccurate eyewitnesses. However, these cues are not powerful enough
to allow laypersons to differentiate between accurate and inaccurate testimony,
290
Handbook of Psychology of Investigative Interviewing
or between truthful and deceptive eyewitnesses, beyond what would be
expected by chance.
Deceptive s uspects ’ e vent r ecall and d uration e stimates
a cross r epeated i nterviews
Consider the following scenario: Two men, John Brown and Joe Smith, inde-
pendently fi nd an unmarked envelope in a city street. Brown ’ s envelope con-
tains $25,000 in bank notes payable to the holder, whereas Smith ’ s envelope
contains some worthless advertisements. Brown realizes that the envelope has
been lost and the proper thing to do would be to fi nd its rightful owner by
calling the police. Instead, he makes the decision to keep the money, knowing
full well that, from this point onward, he will have to lie to protect himself if
he is ever questioned about how he obtained this money. A day later the police
contact both Brown and Smith
because they were reportedly seen in the area
where the money had gone missing. They are repeatedly questioned about
their activities during the general time period when the bank notes were lost.
Clearly, the two men have different motivations in describing their activities
to the authorities. John Brown wants to keep the money and has decided to
be deceptive in his reports, whereas Joe Smith, who has done nothing wrong,
attempts to be accurate and truthful in his reports.
The research to be described in this section mimics the above scenario
(Yarmey, Wells & Yuval, 2006 ). This experiment had two purposes: fi rst, to
explore the differences in reports of deceptive suspects and truthful suspects
in their repeated accounts of personal activities, including a staged crime or a
staged non - criminal incident, respectively; and second, to determine if time
estimations may be used as an indicator of deception.
Truthful and deceptive persons do differ somewhat in their verbal reports.
For example, in contrast to truth - tellers, liars tend to use fewer words and
offer fewer details in their stories (Undeutsch, 1982 ; deTurck & Miller, 1985 ;
Zuckerman & Driver, 1985 ; Miller & Stiff, 1993 ; Anolli & Ciceri, 1997 ;
Yuval, 2003 ). This may be because of the ease with which fewer details can
be stored and remembered when liars are asked to repeat their testimony (Vrij,
2000 ). Compared to truth - tellers, the stories reported by liars have been found
to be less plausible but more structured in the chronological order of story -
telling (i.e., this happened fi rst followed by this, then that, and so forth)
(Zaparniuk, Yuille & Taylor, 1995 ; Granhag & Vrij, 2005 ).
Granhag
& Str
ö mwall
(2000) suggest that the
‘ repeat vs. reconstruct
’
hypothesis can account for the different strategies that liars and truth - tellers
employ over repeated interviews. They argue that deceivers recognize that, in
order to avoid detection, they have to retain a clear memory of what they said
in previous accounts. Second, because memory is reconstructive (Loftus, 1979 ;
Baddeley,
1990 ), individuals who are telling the truth will gain, lose and
change information over time. Thus, deceivers will attempt to repeat what they
Handbook of Psychology of Investigative Interviewing: Current Developments and Future Directions Page 55