cated these fi ndings with four and fi ve year olds. Importantly, their research
established that just two CI instructions (
contest reinstatement and
report
everything ) in combination reduced children
’ s reporting of misinformation
during the interview and later memory tests. We shall return to this study
below.
Do CI s e nhance o lder w itnesses ’ r ecollections?
Older adults represent a special group of witnesses. In many countries, growing
numbers of older adults are remaining active in the community. Hence, it is
likely that some will witness or be a victim of a crime. Moreover, physical and
The Cognitive Interview
143
emotional abuse of the elderly is being reported with increasing frequency
(Action on Elder Abuse, 2004 ) with those over the age of 75 years being
particularly vulnerable. Obtaining reliable eyewitness testimony from older
witnesses has now become a key concern to policy - makers and professionals,
yet only a small number of laboratory studies have evaluated the reliability
and accuracy of older adults ’ recollections in an eyewitness context. Findings
from these studies reveal that memory recall is less complete and less accu-
rate in comparison to young adults, whether the witnessed event is a slide
show (Yarmey & Kent, 1980 ), a short fi lm (List, 1986 ) or a staged event
(Yarmey, 1993 ). Such age differences are reported when memory is tested
immediately, minutes or days after the witnessed event (Brimacombe, Quinton,
Nance & Garrioch, 1997 ; see Mueller - Johnson & Ceci, 2007 , for an excellent
review).
A handful of studies have evaluated the effectiveness of CI techniques with
older adults. In an early study, young (18 – 35 years) and old (65 – 80 years)
adults were shown a short fi lm of a staged robbery (Mello & Fisher, 1996 ).
Half an hour later, both groups were given a regular police interview, a full
CI or a modifi ed CI (omitted change perspective ). An unexpected age pattern
was found in that when given a full CI older adults provided more correct
information than young adults. Research by McMahon
(2000) evaluated
whether a full CI would increase correct recall of a fi lm of a simulated crime.
Young (18 – 50 years) and old (60 – 88 years) adults were interviewed 30 minutes
after viewing the fi lm. As expected, the younger adults recollected more correct
information than the older adults, but only in the control interview (see also
Isingrini, Vazou & Leroy, 1995 ). In Rose, Bull & Vrij ’ s (2003) study, young
(18 – 31 years) and old (59 – 84 years) adults viewed a short fi lm of a staged
robbery. Thirty minutes later all participants were presented with line - up iden-
tifi cation tasks. Whilst the older adults made more incorrect identifi cations
than the young adults, no effects of mental or physical context reinstatement
instructions on performance were found. In a second recent line - up study with
young (16 – 30 years) and old (64 – 86 years) adults, Wilcock, Bull & Vrij (2007)
reported that context reinstatement increased correct rejections in target
absent line - ups but only in the old adults.
More recently, two studies by Wright and Holliday provided evidence that
CI protocols can increase correct recollections in elderly witnesses. Wright &
Holliday (2007a) evaluated older witnesses ’ recall of a short fi lm using full CI,
a MCI (omitting change perspective instruction) or a control interview. The
full CI increased correct recall by 20% for young adults (aged 17 – 31 years),
27% for young - old adults (aged 60 – 74 years), and 18% for old - old adults (aged
75 – 95 years), while the MCI increased correct recall by 14% for young adults,
17% for young - old adults and 15% for old - old adults. In the second study,
Wright
& Holliday
(2007b) compared the recollections of old adults
(aged 75 – 96 years) who displayed evidence of cognitive impairments on the
Mini - Mental State Examination (MMSE) (Folstein, Folstein
& McHugh,
144
Handbook of Psychology of Investigative Interviewing
1975 ) with old adults without cognitive impairments on MMSE. Old adults
with a low MMSE score recalled fewer correct details and were less accu-
rate than those with high scores. Importantly, high and low MMSE old adult
groups reported substantially more correct information about Action, Person,
Object and Surrounding details with a MCI than with a control interview.
Do CI t echniques i mprove i ntellectually d isabled
w itnesses ’ r ecollections?
Individuals with intellectual disabilities (ID) are considered vulnerable
witnesses. Research with these adults reports that they are slower than
typically developed adults to encode, store and retrieve details of an event
(Milne & Bull, 2001 ). Adults with IDs have been considered by courts to be
unreliable witnesses ( ibid. ). Yet the information that ID adults do recollect is
just as accurate as that of other adult witnesses. However, ID adults do report
fewer details of a witnessed event than other adults (Perlman, Ericsson, Esses
& Isaacs, 1994 ; Milne & Bull, 2001 ). Given that these witnesses have been
found to be highly suggestible (Cardone & Dent, 1996 ; Milne, Clare & Bull,
1999 ), it is of major importance that ID adults, like other groups of vulnerable
witnesses, be questioned appropriately and non
- suggestively. Research has
demonstrated that ID adults are particularly susceptible to the negative effects
of social demand factors. Kebbell & Hatton (1999) reviewed the research
literature and reported that ID adults are likely to say ‘ yes ’ to questions irre-
spective of the content of such questions. ID adults are more likely than other
adults to fabricate and change answers in response to the interviewer ’ ques-
tions (Clare & Gudjonsson, 1993 ; Ternes & Yuille, 2008 ). They are also
highly suggestible (Milne, Clare & Bull, 2002 ; Ternes & Yuille, 2008 ).
The body of research in which CI protocols have been evaluated with cog-
nitively impaired adults is sparse. In the fi rst such study, Brown & Geiselman
(1990) reported that ID adults recalled fewer correct details than other adults.
This effect did not vary by type of interview (whether CI or control). ID adults
also reported more confabulated details during a CI than other adults. In
Milne, Clare & Bull ’ s (1999) study, adults with and without mild ID watched
a short fi lm of an accident and were interviewed the next day with either a CI
or a structured (control) interview (K ö hnken, 1993 ). An increase in correct
details was found for those given a CI. However, ID adults given a CI reported
more confabulated details than ID adults given a control interview (Milne &
Bull, 2001 ). More recently, CI principles have been tested on elderly adults
with dementia (Wright & Holliday, 2007b ). Adults (75 – 96 years) with and
without cognitive impairments were given a modifi ed CI (omit CP), or a full
CI, or a control interview following viewing a short fi lm. Cognitive - impaired
adults remembered fewer c
orrect details than non - impaired adults. Nonetheless,
both types of CIs enhanced recollections of both impaired and non - impaired
elderly adults, although impaired participants had particular diffi culties with
The Cognitive Interview
145
the CP instruction. Clearly, more research is needed before fi rm conclusions
can be drawn about the benefi ts (or not) of using CI protocols with these
vulnerable individuals.
The research literature in which CI protocols have been evaluated with
intellectually disabled (ID) children is extremely small. At the time of writing,
all this work has been conducted in the UK by Milne and Bull. Several years
ago, Milne & Bull (2001) pointed out in a review article that the recollections
of ID children for events they have experienced or witnessed are typically less
complete than those of children without ID (much like ID adults). Therefore,
they argued ( ibid .) that it is crucial that researchers evaluate protocols that
have the potential to enhance witness testimony, such as the CI. Milne & Bull
(1996) gave ID children a CI or a control interview similar to the MOGP
interview (Home Offi ce, 1992 ) following a witnessed event. Children given a
CI recollected a greater number of correct details in comparison to children
given a control interview. Signifi cantly, a CI did not lead to increases in report-
ing of incorrect and confabulated details.
How e ffective a re CI p rotocols in the fi eld?
The revised CI has proved benefi cial when evaluated with adult witnesses and
victims of real crime. In the fi rst study (Fisher, Geiselman & Amador, 1989 ),
police detectives from the Metro - Dade Police Department, Miami conducted
interviews before and after four hours ’ training in the revised CI protocols.
Importantly, the amount of information obtained from witnesses by police
after training increased substantially (47%). A second fi eld study was conducted
in the UK by George (1991, cited in George & Clifford, 1996 ). Police who
had been trained in the CI protocols interviewed young adults after they had
witnessed a staged argument. Recall of correct details was substantially higher
in CIs than in control interviews, without a concomitant increase in reporting
of incorrect information. Notably, the size of this recall advantage is remark-
ably similar to fi ndings from laboratory studies (Fisher & Schreiber, 2007 ).
More recently, Fisher & Castano (2008) reported that CI protocols have been
used effectively in the USA in several police investigations, including child
sexual abuse, kidnapping, a bombing and a murder enquiry. In the last case,
a CI was used with a female who had witnessed a murder 33 years earlier when
she was fi ve years old.
Promising a pplications of the CI p rotocols
Do the a dvantages of a CI r emain with a s hortened v ersion?
For practitioners, shortened interview protocols that facilitate accurate recol-
lections are important given the potential problems surrounding interviewing
146
Handbook of Psychology of Investigative Interviewing
vulnerable witnesses (limited attention span, shorter memory) and pressures
on police and other professionals to obtain maximum information as soon
as possible after a crime. Wright & Holliday (2005) conducted a study in
which police offi cers from a number of constabularies in the UK responded
to a questionnaire. Police who were reluctant to use CI techniques stated that
such protocols are far too long and mentally demanding to be of practical use
when interviewing elderly witnesses and victims. Davis, McMahon
&
Greenwood
(2005) evaluated a shortened version of a CI (CRI and RE
instructions). College students viewed a short fi lm of a staged crime and were
subsequently interviewed with a full CI, a short CI or a control interview.
Importantly, correct recollection of the fi lm details was higher in the full CI
and the short CI (equally) than in a control interview condition. Dando,
Wilcock & Milne (2009) , using college students, investigated two variations
of context reinstatement instructions: mental context reinstatement (MRI) as
used in cognitive interviews, and a shortened form they called ‘ sketch rein-
statement of context ’ (MRC). The sketch MRC condition produced more
accurate recollections of a fi lmed crime than the MRI and control interview
conditions.
Holliday & Albon (2004) conducted a comprehensive study that aimed
to develop a developmentally appropriate shortened CI for young children
which minimized suggestibility. The rationale behind this research was
Holliday ’ s (2003b) fi ndings that a Modifi ed CI (MCI) reduced children ’ s
acceptance and subsequent reporting of suggestions. Given the short attention
span of very young children, a shortened version of the MCI would be par-
ticularly useful. Children viewed a short fi lm followed by exposure to a number
of misleading suggestions. Children were then administered one of six inter-
view protocols:
• a control interview (structured interview; K ö hnken, 1993 );
• a full CI (FCI) containing the CRI, RE, CO and CP instructions;
• a Modifi ed CI, which was identical to the FCI except CP was omitted;
• an Enhanced Rapport MCI (ERMCI), which was identical to the MCI
except for an additional fi ve minutes of rapport - building in which children
described a favourite game;
• a RE and CO interview, which was the same as the MCI except that CRI
was omitted;
• a RE and CRI interview which resembled the MCI except that CO was
omitted.
These interview protocols are presented in Table 9.1 .
In terms of the quality of communication (e.g. rapport
- building, active
listening) and questioning methods, SIs and CIs used in Holliday & Albon ’ s
(2004) research are identical
– each interview protocol employs a phased
The Cognitive Interview
147
Table 9.1: Child interview protocols tested in Holliday & Albon ’ s (2004) research
Cognitive interviews
Control interview
1. Rapport - building phase
1. Rapport - building phase
Chat about general interests the child
Chat about General interests of
(e.g., pets, football)
the child (e.g. of pets, football)
Describe their favourite game 1
Explain aims/rules of interview a
Explain aims/rules of interview a
2. Free recall phase
2. Free recall phase
a. Context reinstatement 1,2,3,5
b. Report everything 1,2,3,4,5
c. Change order 1,2,3,4
d. Change perspective 3
e. Free recall report request a
Free recall report request a
3. Questioning phase
3. Questioning phase
Details reported in Free Recall Phase
Details reported in Free Recall
are used as the bases of open - ended
Phase are used as the bases of
and specifi c and questions in this
open - ended specifi c questions in
phase. a
thi
s phase. a
4. Closure a
4. Closure a
Notes :
a All interviews.
1. Enhanced rapport CI. 2. MCI. 3. Full CI. 4. Report everything and change order interview.
5. Report everything and context reinstatement interview.
approach proceeding from free recall to open, to closed, to specifi c
questions.
The CIs produced more correct recollections than a control interview.
As has been reported previously, CIs enhanced children
’ s recall of
person, action and object information (cf. Milne & Bull, 2002; 2003 ; Holli-
day,
2003a, 2003b
), without an accompanying increase in reporting of
incorrect or confabulated details (cf. McCauley & Fisher, 1996 ; Granhag &
Spjut,
2001 ; Holliday,
2003a; 2003b
). The fi nding that young children
’ s
recall of person details can be improved with CI instructions is very important
given that investigative interviews necessarily require accurate witness
descriptions.
In earlier CI research, concerns were raised that the CO and CP mnemon-
ics might encourage young children to confabulate (e.g. Ceci, Bruck & Battin,
2000 ; Memon & K ö hnken, 1992 ; Saywitz et al ., 1992 ), although others (e.g.
Milne & Bull, 2002 ; Holliday, 2003a; 2003b ) have reported no increase in
148
Handbook of Psychology of Investigative Interviewing
reporting of incorrect or confabulated details in young children. The CO
mnemonic was included in MCI, FCI, Enhanced Rapport Modifi ed Cognitive
Interview (ERMCI), and RE and CO interviews because research has shown
that children can manage this instruction when it is accompanied by frequent
prompts (Milne & Bull, 2002 ; Holliday, 2003a; 2003b ). Given the concerns
that correct implementation of the CP instruction may be beyond the capabili-
ties of very young children (Geiselman
& Padilla,
1988 ; Newcombe
&
Huttenlocher, 1992 ), and that it is unpopular with police offi cers in England
(Memon & Stevenage, 1996 ), this instruction was only included in a FCI with
the specifi c aim of evaluating its effectiveness with young children. However,
as reported by others (e.g. Milne & Bull, 2002 ; Holliday, 2003a; 2003b )
children had little diffi culty in using the CO mnemonic, demonstrating the
Handbook of Psychology of Investigative Interviewing: Current Developments and Future Directions Page 28
