Memon, A., Wark, L., Holley, A., Bull, R. & Koehnken, G. (1997) Eyewitness performance in Cognitive and Structured Interviews. Memory, 5, 639-655.

Eyewitness performance in Cognitive and Structured Interviews

Amina Memon[1],
University of Texas at Dallas,
School of Human Development,
TX 75083-0688,

Linsey Wark & Angela Holley
Department of Psychology
University of Southampton
S017 1BJ

Ray Bull
Department of Psychology
University of Portsmouth
P01 2ER

Guenter Koehnken
Institut fur psychologie
Universitat Kiel
Olshausenstr 40-60
24098 Kiel

Keywords: cognitive interview, eyewitness, retrieval, mnemonics, training


This paper addresses two methodological and theoretical questions relating to the Cognitive Interview (CI), which previous research has found to increase witness recall in interviews. (1) To what extent are the effects of the CI mnemonic techniques when communication techniques are held constant? (2) How do trained interviewers compare with untrained interviewers? In this study, witnesses (college students) viewed a short film clip of a shooting and were questioned by interviewers (research assistants) trained in conducting the CI or a Structured Interview (SI), similar to the CI save for the `cognitive' components, or by untrained interviewers (UI). The CI and SI groups recalled significantly more correct information compared to the UI group. However they also reported more errors and confabulated details. Theoretical and practical implications of the results are discussed in terms of precisely identifying the CI facilitatory effects and consequent good practice in the forensic setting.

Eyewitness Performance in Cognitive and Structured Interviews

The Cognitive Interview (CI) is a technique which has undergone considerable modification in the last ten years designed to optimise its effectiveness in enhancing eyewitness recall (Fisher, Geiselman & Amador, 1989; Fisher & Geiselman, 1992; Memon & Koehnken, 1992; Koehnken, 1995; Memon, 1996). In an extension of the CI, Fisher and Geiselman (1992) recommend a number of additional techniques be used (we shall refer to this version as the enhanced CI). The enhanced CI has not been extensively tested, most studies rely on the original procedure. The present study tests the enhanced CI as described in Fisher and Geiselman (1992) against two controls. The first control is an interview procedure which incorporates some of the communication elements of the enhanced CI (but not its original cognitive components) and the second is an untrained control.

The Cognitive interview

In its original form the CI comprised several memory retrieval techniques together with some supplementary techniques for recalling specific details (Geiselman, 1994). The original CI cognitive techniques drew upon the theoretical principle that there are several retrieval paths to memory for an event and information not accessible with one technique may be accessible with another (Tulving, 1974). One of the main CI techniques is mentally to reconstruct the physical and personal contexts that existed at the time of the witnessed event. This is based upon the assumption that context reinstatement increases the accessibility of stored information (Tulving & Thomson's Encoding Specificity Hypothesis, 1973). Context reinstatement is achieved with various instructions from the interviewer to the witness. For example, the witness is encouraged to recreate mentally the environmental aspects of the original scene, to comment on his/her emotional reactions and feelings at the time, and to describe any sounds, smells and physical conditions. The instruction to reinstate context is administered prior to eliciting a narrative account of the event (i.e. before free report). A further context reinstatement instruction is administered prior to the questioning part of the interview in order to obtain more specific details (see Memon, Wark, Bull and Koehnken, in press for examples). Fisher and Geiselman (1992) suggest that more recall details can be elicited by activating and probing a witness's mental image of the various parts of an event, such as a suspect's face, clothing, objects. With reference to the dual coding hypothesis (Paivio, 1971) a distinction is drawn between conceptual image codes (an image stored as a concept or dictionary definition) and pictorial codes (the mental representation of an image). The latter are believed to elicit more detailed information than the former. For example, in retrieving information about a face using the concept code, one may be able to access information that the face is disfigured. If then asked to form an visual image of the face a witness may be able to give more details of this, such as the presence of a scar across the cheek.

Another CI technique is to ask the witness to `report everything'. For instance, witnesses are encouraged to report in full without screening out anything they consider to be irrelevant or for which they have only partial recall (Fisher & Geiselman, 1992).

There are two additional CI techniques-the instruction to recall the event in different orders (i.e. to recall from the beginning, the end and/or the most salient detail) and to recall from different perspectives- from the perspective of another person or by placing oneself in a different location- that form part of the original procedure. These techniques may work, in part, by reducing the extent to which prior knowledge, expectations and schema influence and/or limit recall (Norman & Bobrow, 1978). However, there are some concerns about using these techniques, such as the possibility of interviewees misunderstanding the instructions as a request for inferences or speculations (see Memon, Cronin, Eaves & Bull, 1993; Memon & Koehnken, 1992). Moreover, police officers are reluctant to use these techniques for these reasons (George, 1992) and concerns about presenting evidence arising from the use of this technique in court. Finally, previous research has shown that the reverse order recall instruction has no benefits over an instruction to merely make one additional retrieval attempt (Memon et al, in press). The present study will not include these elements of the original CI in the training of interviewers.

Following the early success of the original CI procedure, a content analysis of real police interviews was undertaken (Fisher, Geiselman & Raymond, 1987a) and deficiencies in communication skills were noted. The enhanced CI was an important development as it was specifically directed towards addressing problems with the typical police interview. The current paper will concentrate on the enhanced version of the CI since it is the one recommended to professionals (Fisher & Geiselman, 1992).

The enhanced CI utilises the techniques of the original CI. In addition to these specific strategies, an interviewer trained in the use of the enhanced CI attempts to facilitate interviewer-witness communication by building rapport with the witness and by transferring control to the witness. The latter is achieved by (a) making it clear to the witness that the interviewer does not have knowledge of the event and that the witness has to do the work; (b) allowing the witness time to think and respond, and (c) timing questions to suit the witness. The rapport building, together with transfer of control, is thought to increase the flow of information (Fisher & Geiselman, 1992; Memon, 1996). Hence any increase in amount of information reported may partly be a consequence of communication and rapport and not solely of cognitive memory enhancement. Indeed in the counselling domain the association of rapport and transfer of control with good interviewing practice is nothing new. In 1965, Richardson, Dohrenwood and Klein wrote:

`Rather than impose his questioning framework on the respondent, the experienced interviewer adapts his pace, idiom and thought to those of the respondent.' (p. 285).

The enhanced CI is essentially a non-directive interview (Roethlibergert & Dickson, 1946) and its `communication' components resemble techniques used by Carl Rogers (1942) in counselling and psychotherapy. An important question, therefore, is whether the success of the enhanced CI is due more to the use of the purposely included non-directive interview techniques (such as those described immediately above) than to its cognitive `memory retrieval' techniques.

So what evidence is there to suggest the CI is an effective procedure to use when interviewing witnesses? In early studies which examined the effectiveness of the original CI (e.g., Geiselman, Fisher, Mackinnon & Holland, 1985) it was compared with a "standard" interview or the techniques police officers typically use (see Fisher, Geiselman and Raymond, 1987a)[2]. This showed that with relatively brief training students and police officers could elicit significantly more correct information from witnesses (college students and non-college samples) without influencing the error rate. Tests of the enhanced CI found that it could generate even more information than the original CI (Fisher, Geiselman, Raymond, Jurkevich and Warhaftig, 1987b). Again, the numbers of errors that were elicited and the proportion of correct and incorrect information (accuracy) was unaffected. In light of these results, the enhanced CI was presented as a reliable repertoire of techniques to maximise the recall of a witness (Fisher & Geiselman, 1992). More recent attempts at replicating the earlier studies have identified a number of problems. The CI does not always facilitate recall when compared to a structured interview (Memon, Cronin, Eaves & Bull, 1993; Memon, Cronin, Eaves and Bull, 1996a) or a recognition checklist (Cohen and Java, 1995). When CI training is ineffective, trained interviewers may fare worse than an untrained group (Memon, Milne, Koehnken & Bull, 1994). Furthermore, recent tests of the CI which have consistently shown that while the CI and control groups do not differ in accuracy rates there is an increase in the number of errors generated in the CI condition (Bekerian, Dennett, Reeder, Slopper, Saunders & Evans, 1994; Koehnken, Finger, Nitschke, Hoefer & Aschermann, 1992; Memon et al, in press; Milne, Bull, Koehnken & Memon, 1995; McCauley & Fisher, 1995; Mantwill, Koehnken & Aschermann, 1995). Further laboratory research on the CI is therefore needed in order to understand the conditions under which the CI will be effective and when it may not.

From a practical perspective, it is important to show that the CI is more effective than the techniques currently in use by police officers and others as Fisher and Geiselman have done. From a theoretical perspective, a proper experimental control group is needed to demonstrate that the CI techniques themselves are causing the effects and not some artifactual side effect arising from the training or quality of the CI interviewers. In the present research we sought to construct a control condition (the structured interview or SI) in which the quality of training in communication and questioning techniques was comparable to the CI condition. The structure of this training for our CI and SI followed that recommended to professionals who interview witnesses, both adults (Koehnken, 1995) and children (the 1992 Government recommended Memorandum of Good Practice, see Bull, 1992; 1995). The essence of these recommendations is to treat the interview as a procedure in which a variety of interviewing techniques (including components of the structured interview) are used in relatively discrete phases proceeding from free recall to open questions to closed questions. Thus the structured interview is a suitable control. Moreover, the CI versus SI comparison in the present study lets one examine the effect of mnemonic techniques (communication techniques are held constant). The SI versus untrained will enable us to look at the effect of communication techniques.

With respect to the increases in errors, further research is needed to understand the conditions in which these errors are likely to occur so as they may be prevented. It should be noted that the increases in errors is relatively small when compared to the large gains in correct details (see Memon et al in press). Nevertheless, errors are obviously of great concern from a forensic perspective. Research is required in order to understand where in the interview they occur, whether they are associated with the use of particular techniques, and the types of information that is likely to be erroneously recalled. Based upon our earlier research (Memon et al, in press; Milne et al, 1995) we predict that the errors will occur in the questioning phase. The questioning phase in the CI is characterised by the use of imagery probes and detailed questioning. The increase in errors could be a result of instructions to elicit images in detail (Bekerian, Dennett, Hill & Hitchcock, 1990), source misattributions due to similarities between imagined and perceptual details (Carris, 1992), and /or to the use of specific questions. Data from the SI group (specific questions without imagery) will facilitate the further examination of these hypotheses.

The aims of this study are therefore two fold. The first is to determine the extent to which the effects of the enhanced CI may be a result of improved rapport and communication between the interviewer and witness. The second is to consider the possible aspects of the enhanced CI which may give rise to the reporting of errors. The enhanced CI is compared with a structured Interview (SI) and untrained control (UI) group.



The effects of interview condition (Cognitive, Structured or Untrained) on total correct recall, errors and confabulations were examined using a between-subjects design.


Data were collected from sixty-six participants (forty female and twenty-six male) who were `sixth form' students at a local college. They were aged between 16 and 19 years of age (mean=16 years, 7 months) and were randomly assigned to the experimental conditions. Previous studies of the CI have tended to use older age groups (e.g Geiselman et al, 1985) or children (e.g. McCauley & Fisher, 1995). Adolescents are just as likely to serve as witnesses in the criminal justice system and therefore this age group was selected for the present study.

Interviewer Training

The interviewers were all female research assistants[3] who received payment for the interviews conducted. All interviewers were university students in Psychology. There were three interviewers in each condition. The CI and SI groups were trained in two 4-hour sessions which began with an introductory lecture on the importance of the interview in psychological assessment and information gathering. Both groups were given guidelines about non-verbal behaviour in the interview (e.g., seating position, eye contact, pauses and speech rate). (There was an opportunity to practise and receive feedback on these later in the training). Both groups were then informed that each interview should comprise a number of phases. For each phase there was a demonstration role-play which was followed by a live practice role-play (interviewers were asked to choose an event, playing the part of interviewee and interviewer) these latter being video-recorded. The interviewers then received feedback on their interviews. There were plenary discussion/question sessions at the end of the training. In addition to active role-plays, the interviewers were encouraged to rehearse mentally the various phases of the interview. For each group (i.e., CI and SI) the interview was divided into the phases of free recall (FR) and questioning (QU). Prior to this both CI and SI interviewers engaged in building rapport with the witness by asking open-ended questions (not about the to be recalled event) to generate casual conversation. They also used this time to make it clear that they did not have any knowledge of the event (this was true) and that the witness could dictate the pace of the interview. The last two instructions were part of the `transfer of control' procedure. The FR and QU procedures were as follows:

In the free recall phase the SI interviewers and the CI interviewers were asked to request a free narrative account from the witness and to use this as a strategy for obtaining information to be referred to later in the questioning phase. The CI interviewers were also given training in encouraging witnesses to reinstate mentally the context (as described in the introduction) prior to obtaining a free report. The CI interviewers also incorporated at this stage the instruction to report everything.

In the questioning phase, the CI and SI interviewers were asked to use the information reported in free recall as a guide for follow-up questions. (Both the CI and SI interviewers had also been similarly instructed in the use of appropriate types of questions). They were asked to begin with open questions and then follow these with closed questions. In general interviewers were asked to use the QU phase to find out who was present at the event and what they did. Where a person was mentioned, interviewers were asked to elicit details about clothing. They were specifically instructed to avoid leading, misleading, and forced-choice questions. The CI interviewers received additional training in the activation and probing of images relating to various parts of the event.

The untrained interviewers (n=3) received no prior training or information on the techniques to be used. They received a brief protocol asking them to generate a free report and then ask questions but were given no instructions on the types of questions to ask or interview strategy. They were merely asked to find out about the witnessed event. All interviewers were asked to do their best and get as much information as possible.

The interviewer protocols are presented in Appendix I.


Participant witnesses were asked to take part in a study on memory and were shown a short video clip (1.5 minutes in duration) of a shooting incident in which a young boy is murdered. The event was shown to participants in groups of four to six. Witnesses were interviewed individually five days after the event by a CI, SI or untrained interviewer.

Coding and scoring of the interview transcripts.

The interviews were transcribed verbatim and the resultant details coded and scored. The transcribed coding will be discussed in two ways involving (i) information recalled and (ii) interviewer technique.

The participants' recall was classified as either correct, as error (e.g. saying red for the colour of the silver car which was present) or as confabulation (mentioning a detail that was not present). Given that the difference between errors and confabulations is important in a forensic setting (Gudjonsson and Clare, in press), separate analyses of errors and confabulations were undertaken. A confabulation about a person (e.g. describing a person who was not present at the event in question) may have quite different ramifications for an investigation than a person error (describing a person's jacket as black when in the actual event it had been blue).

In order to score the transcripts, two research assistants made a detailed transcription of the video clip. At this stage ten transcripts (picked at random) were examined and the detailed coding scheme was expanded to include any items mentioned by the interviewees which could be confirmed from the video. These details were classified as either pertaining to people (person details), actions (action details), objects in the video (object details) or to the environment in which the scenes took place (location details). This procedure enabled us to come up with an exhaustive list of details based on what witnesses recalled than relying upon an experimenter generated summary of the event. The final coding scheme comprised a total of 178 units of information which were part of the event.

The position the information occupied in the interview (FR or QU) was noted and the data from each stage were examined separately. After a participant's free report was scored, information appearing in the questioning phase of the interview was only scored if it was new. Given that information at each phase of the interview was related to what was reported earlier, interview phase was not treated as a repeated measures factor in the analyses. Two research assistants coded and scored the interview transcripts. In addition ten transcripts (15% of the total) were coded independently by both of the research assistants and compared. The following correlations of the two coders' scores were found: total accurate, r=.99 p<.0001, total error, r=.90 p<.0001 and total confabulations, r=.94 p<.0001.

Number of questions.

A count was taken of the number of questions asked in the QU phase. Questions which were repeated (if the participant had not understood the question initially) were only scored once and questions clarifying a piece of information given by the participant were not scored (e.g. "So the gun was silver?" when the participant had just provided this information). Where the interviewer was using imagery instructions to focus the interviewee on a particular area (e.g. to describe the gunman's face) these instructions were coded separately as "CI probes". These were often used by some of the interviewers in conjunction with open questions and follow-up questions (closed questions). The following example would be scored as 1 CI probe (the gunman's face), one open question. "Now I want you to concentrate and picture the gunman's face in your mind. You might prefer to close your eyes. Have you got a clear image of the gunman's face? Can you describe the gunman's face?". A closed question was used to get more specific information when a partial description had already been elicited from a prior open question.

The number of CI probes used in the questioning phase of each interview (especially CI) was tallied. (None of the SI or untrained interviews were found to use imagery or context reinstatement spontaneously.) The use of the other instructions (for example, report everything, transfer control) were noted as either being present or absent from the interviews.


(1) Recall Data

A series of one way analyses of variance (ANOVA) was performed on the recall data in order to compare performance across CI (n=22), SI (n=23) and UI (n=21) conditions[4]. The dependent measures were the total number of correct details, errors and confabulations elicited per interviewee. The amount of correct, incorrect and confabulated details will be reported first followed by the proportion of recall that was correct (i.e. accuracy).

(insert table one about here)

(a) Total amount recalled.

As predicted from earlier research on the CI, there was a significant difference between conditions in the total number of correct details recalled (F (2,63)=4.30, p<.05). Post-hoc analyses showed that significantly more correct information was elicited in the CI condition than the UI condition (Fisher PLSD= 11.21, p<.01). Contrary to expectation, however, there was no significant difference between CI and SI condition (Fisher PLSD= 7.19, p=.07). There was also no significant difference between the SI and UI (Fisher PLSD= 3.28, p>.05).

(b) Errors.

The errors followed a similar pattern. There was a significant effect of interview condition (F (2,63)=6.27, p<.01). Post-hoc comparisons showed more errors were produced in the CI condition as compared to the UI (Fisher PLSD= 3.08, p<.001) and more errors in the SI as compared to the UI (Fisher PLSD= 2.01, p<.05). Again the CI and SI did not differ significantly Fisher PLSD= 1.06, p>.05).

(c) Confabulations.

Finally, there was a significant effect of interview condition on confabulations (F (2,63)= 6.14, p<.01). As with the errors, there were more confabulations in the CI condition as compared to the UI (Fisher PLSD= 3.08, p<.01) and more confabulations in the SI as compared to the UI (Fisher PLSD= 2.72, p<.01). There were no significant differences between the CI and SI (p>.05).

(d) Accuracy rates.

The accuracy rates (i.e. number of correct details divided by total number of details) for the CI, SI and untrained group were 84%, 84% and 92% respectively. There was a significant difference between the groups in these accuracy rates (F (2, 63)= 9.85, p<.001). Post-hoc analyses showed the CI and SI interviews were significantly less accurate than the UI interviews (Fisher PLSD= .076, p<.001 and Fisher PLSD= .079, p<.00 respectively) but not different from each other.

To summarise the results so far, performance in the CI and SI conditions did not differ significantly from each other in terms of the number of correct details, errors and confabulations. The CI and SI both produced more errors and confabulations than the UI, and the CI more correct recall than the UI. This is the first study conducted in our laboratory which shows an increase in confabulations with the CI, although a similar effect has been noted in a recent study conducted in a different laboratory (Gwyer, Clifford & Dritschel, 1995). The next section compares performance across interview categories under free recall conditions (see table 1).

(e) Recall in different phases.

In the free recall phase there were no significant differences between conditions in total correct details or errors (Fs <1). There were however significant differences in the total number of confabulations (F (2,63) = 5.75, p<.01) with more confabulations in the CI as compared to the UI (Fisher PLSD = .72, p<.01) and in the SI as compared to the UI (Fisher PLSD=.74, p<.01). The CI and SI were not significantly different. This was unexpected and surprising as there were no questions asked during the free recall phase. This argues that the increased number of confabulations cannot be due to (inappropriate) questioning. It should be pointed out however, that accuracy rates were high in free recall and the number of confabulations was relatively small (approximately 0.5-1.5 confabulations for every 30 units of correct information).

In the questioning phase there was a significant effect of interview condition for total correct details (F (2, 63) = 5.35, p<.01). Post-hoc analyses showed significantly more correct information in the CI as compared to the UI group (Fisher PLSD= 4.32, p<.01). The post-hoc test did not reveal a significant difference between the CI and SI conditions or SI and UI conditions (p>.05). Looking at the number of errors in the QU phase, there was again significant effect of interview condition (F (2, 63)= 8.09, p<.01). Post-hoc analyses showed more errors in the CI as compared to the UI (Fisher PLSD= 1.44, p<.01) and in the SI as compared to the UI (Fisher PLSD=1.43, p<.01). CI and SI did not differ significantly. There was also an effect for confabulations (F (2, 63)= 5.07, p<.01) with more confabulations in the CI as compared to the UI (Fisher PLSD= 1.67, p<.05) and in the SI as compared to the UI (Fisher PLSD=1.65, p<.01). Again, the CI and SI were not different (p>.05). The increase in confabulations in the CI and SI as compared to the UI were unexpected (cf. Memon, Wark, Holley, Bull and Koehnken, 1996b).

(f) Information type.

Breaking down the total correct details into person, action, objects and location information, there were some significant differences between groups in person details. There were significantly more total person items correctly reported in the CI group as compared to the untrained group (F (2, 63)= 4.01, p<.05, Fisher PLSD= 3.91, p<.01). There were no significant CI/ SI differences in the total correct person information. There were significantly more total person errors in the CI group as compared to the UI group (F (2, 63)= 4.01, p<.05, Fisher PLSD= 1.25, p<.01). Again there were no CI vs. SI differences (cf. Memon et al, in press; Milne et al, 1995). Finally, there was an effect of interview type on person confabulations F (2, 63)= 4.81, p<.05). Post-hoc analyses showed a significant difference between CI and UI (Fisher PLSD= 1.34, p<.01) and the SI and UI (Fisher PLSD= 1.33, p<.01) but not the CI and SI (p>.05). These data are presented in Table 2

(insert table 2 about here)

(h) Event knowledge with each interview

Given that the interviewers in this study were naive about the details of the event until the first interview, it is important to check what impact each successive interview had on amount of information gained from witnesses. For each interviewer in each condition, interview position (first interview, second) was correlated with amount of correct information, errors and confabulated details. There were no significant correlations.

Analysis of interview techniques

On the whole, the interviewers used the techniques that were expected of them with all the CI interviewers using the `transfer of control' and `report everything' instruction. Similarly, they all used context reinstatement for the free report and all used CI probes in the questioning phase.

The SI interviewers did not on the whole use the CI instructions except for the `report everything' instruction which was used four times (17%). In accordance with instructions SI interviewers used the `transfer of control' in most (74%) of their interviews but less frequently than did the CI interviewers who used the instruction on every occasion (Chi-square(2) = 47.65, p<.001). The differential use of this instruction in the SI group was due to one of the interviewers who did not transfer control in 60% of her interviews.

The untrained interviewers did not use either the CI or SI techniques with the exception of one interviewer who used the instruction to `report everything' on a couple of occasions.

Number of questions asked

The mean number of questions asked in the CI, SI and UI were 17.73, 15.47 and 7.41 respectively (F (2, 63)= 6.21, p<.01). Post-hoc analyses showed significantly more questions were asked in CI than UI (Fisher PLSD= 4.34, p<.01) and in SI than UI (Fisher PLSD= 4.29, p<.05). There were no significant differences between CI and SI in the number of questions asked. The number of leading and misleading questions was less than 1%.

Interview length

The CI interviews were longer than the SI interviews and CI and SI were in turn longer than the untrained interviewers. This is what we typically find with the enhanced CI and is expected for the following reasons. The CI procedure takes longer to administer, the instructions are more complex, the witness takes longer to respond and most importantly the witness says more. Since the aim of a CI is to elicit more information, it would be natural for the interviews to take longer.


The major aim of this paper was to compare the effects of the Cognitive Interview with an interview procedure (SI) as similar as possible to the CI but not containing the special CI techniques of context reinstatement, imagery and `report everything.' When this was done there was no significant advantage of the CI over this type of SI suggesting that increases in amount reported in the many earlier studies CI may be a consequence increased rapport with the witness rather than the use of cognitive CI techniques.

The similarity in performance of cognitive and structured interviewers is interesting and important as it suggests that when rapport and communication are improved significant gains in information may be achieved. However, the role of the cognitive retrieval aids cannot be ruled out. The means consistently favoured the cognitive over the structured, the data are nevertheless inconsistent with the results of some recent studies (using adults as witnesses) which did find the cognitive group significantly superior to the structured group (Koehnken et al, 1994b; Koehnken et al, 1995; Mantwill et al, 1995). It has been suggested that perhaps our cognitive and structured groups more closely resembled each other than those in previous studies (Fisher, personal communication). It may be the case that the communication and cognitive elements of the CI work in tandem. For example, in order to effectively use the context reinstatement strategy, the witness needs time to concentrate and respond without interruptions from the interviewer. Along similar lines Fisher and McCauley (1995) have suggested that the effects of CI reflect both improved memory search and improved communication (see also Memon & Stevenage, 1996). This provides an illustration of the difficulties of studying memory in real world contexts. An interview is an interaction between two people and memory performance is thus undoubtedly influenced both by the technique used to search memory and rapport with the person who is guiding the retrieval process (see Memon, Wark, Holley, Bull and Koehnken, 1996c for a discussion of the importance of interviewer behaviour). We feel this has brought us to the end of our search for what underlies the success of the cognitive interview. If we were to breakdown the CI (and SI) further to examine what was going on, it would mean sacrificing ecological validity. On the other hand, there is much that remains to be understood about the mechanisms underlying CI effects and the conditions under which the CI may be most effective. It is to this that our discussion will now turn.

The CI and SI elicited more correct recall but also more errors and confabulated details than the UI. One plausible explanation for the increase in both correct and incorrect information with the CI and SI is that it was simply due to the greater number of questions asked by CI and SI interviewers. The CI and SI interviewers began their questioning with open-ended questions but probed for details with closed questions. In contrast, the untrained interviewers asked relatively few questions overall and fewer closed questions. One consequence of asking so many questions is that it places pressure on witnesses to give a response rather than providing them with the option of choosing what to report and withholding details about which they are uncertain (see Koriat & Goldsmith, 1994 for a discussion of the effects of cued versus free recall testing on accuracy). The fact that increases in gains in correct information were found for the same types of information (person details) as the errors suggests that a plausible explanation for the increase in errors is that the CI technique shifts the criterion for responding by influencing confidence and willingness to report information (the theoretical basis is illustrated by signal detection theory as noted by Bekerian and Dennett, 1993). Interestingly, a similar effect (an increase of both correct and incorrect) is sometimes seen with hypnosis and where this occurs it is interpreted in terms of a shift in response criterion (Orne, Soskis, Dinges & Orne, 1984). Krass, Kinoshita and McConkey (1989) argue that under hypnosis, subjects may offer as memories reports they would normally reject on the basis of uncertainty. Asking more questions also means generating more information overall and this may also explain why accuracy rates (the proportion of correctly recalled details) were slightly lower in CI/SI groups. The latter was particularly discouraging as the typical finding in our laboratory and others is that there are no differences in accuracy rates between the CI and SI (see Memon and Stevenage, 1996 for a recent review). An anonymous reviewer has suggested that perhaps the witnesses in the present study were less cautious because they were aware of the artificiality of the situation, they knew their testimony would not have any serious implications. The reviewer suggests that it may be possible to obtain more accurate recall by placing special emphasis on the importance of veridicality of reports. This could be done by repeatedly reminding witnesses not to guess or fabricate responses[5] or by asking for certainty judgements.

One of the aims of this study was to compare the CI (and SI) with an untrained group. It could be argued that our untrained group (university students) are not typical in that they tended to ask fewer questions than our trained group, did not interrupt the witness when they were speaking and did not ask leading questions. In other words, they did not possess the interviewing characteristics of police officers as described by Fisher et al. (1987) and George (1992). They also made fewer errors and confabulations. This could simply be due to the fact that our untrained interviewers were aware that their interviews were being recorded and therefore were cautious in their questioning strategy. This possibility has to be taken into consideration in interpreting our results. In terms of practical implications we cannot recommend that interviewers can effectively collect information without training because our untrained group may not be representative. However, our SI group may provide more information about the likely impact that training in basic communication and rapport building is likely to have. Indeed as practice improves and training is identified as an important issue in police organisations (Cherryman & Bull, 1992) the result is likely to be more effectively conducted interviews which yield more information.


Bekerian, D.A. & Dennett, J.L. (1993). The cognitive interview: Reviving the issues. Applied Cognitive Psychology, 7 (4), 275-298.

Bekerian, D.A., Dennett, J.L., Reeder, C., Slopper, K., Saunders, H., & Evans, L. (1994, April). The influence of the cognitive interview technique on attributions of memory. Paper presented at the Annual Conference of British Psychological Society, Brighton.

Bekerian, D.A., Dennett, J.L., Hill, K., & Hitchcock, R. (1990). Effects of detailed imagery on simulated witness recall. Paper presented at the second European Conference on Law and Psychology, Nuremberg, Germany.

Bull, R. (1995). Innovative techniques for the questioning of child witnesses especially those who are young and those with learning disability. In M. Zaragoza, J.R. Graham, G.C.N. Hall, R. Hirschman & Y.S. Ben-Porath (Eds.), Memory and testimony in the child witness Thousand Oaks, CA: Sage.

Bull, R. (1992). Obtaining evidence expertly: The reliability of interviews with child witnesses. Expert Evidence: The International Digest of Human Behaviour, Science and Law, 1, 3-36.

Carris, M. (1992). Visual imagery and its effect on source misattributions in eyewitness memory. Unpublished honors thesis. Kent State University.

Cherryman, J. & Bull, R. (1995) Investigative interviewing. In F. Leishman, B. Loveday and S. Savage (eds.) Core issues in policing. London: Longman.

Fisher, R.P., & Geiselman, R.E. (1992). Memory enhancing techniques for investigative interviewing: The Cognitive Interview. Springfield III: Charles C. Thomas.

Fisher, R.P., Geiselman, R.E., & Amador, M. (1989). Field test of the cognitive interview: enhancing the recollection of actual victims and witnesses of crime. Journal of Applied Psychology, 74, 722-727.

Fisher, R.P., Geiselman, R.E., & Raymond, D.S. (1987a). Critical analysis of police interviewing techniques. Journal of Police Science and Administration, 15, 177-185.

Fisher, R.P., Geiselman, R.E., Raymond, D.S., Jurkevich, L.M. & Warhaftig, M.L. (1987b). Enhancing eyewitness memory: refining the cognitive interview. Journal of Police Science and Administration, 15, 291-297.

Fisher, R. and McCauley, M. (1995) Improving eyewitness testimony with the cognitive interview. In M. Zaragoza, J.R. Graham, G.C.N. Hall, R. Hirschman & Y.S. Ben-Porath (Eds.), Memory and testimony in the child witness (pp. 141-159). Thousand Oaks, CA: Sage.

Geiselman, R.E. (1994, July). Ten years of cognitive interviewing. Paper presented at the Third Practical Aspects of Memory Conference, Maryland.

Geiselman, R.E., Fisher, R.P., MacKinnon, D.P., & Holland, H.L. (1985). Eyewitness memory enhancement in the police interview: Cognitive retrieval mnemonics versus hypnosis. Journal of Applied Psychology, 70, 401-412.

Geiselman, R.E., Fisher, R.P., MacKinnon, D.P., & Holland, H.L. (1986). Eyewitness memory enhancement in the cognitive interview. American Journal of Psychology, 99, 385-401

George, R.C. (1992). A field evaluation of the cognitive interview. Unpublished Master's Thesis, Polytechnic of East London.

Gudjonsson, G. & Clare, I. C.H. (in press) The relationship between confabulation and intellectual ability, interrogative suggestibility and acquiescence. Personality and Individual Differences.

Gwyer, P., Clifford, B.R. & Dritschel, B. (1995). The effects of the Cognitive Interview on recall, recognition and the confidence/accuracy relationship. Manuscript submitted for publication.

Koehnken, G. (1995). Interviewing adults. In R. Bull & D. Carson (Eds.) Handbook of psychology in legal contexts. Chichester: Wiley.

Koehnken, G., Thurer, C., & Zorberbier, D. (1994b). The cognitive interview: Are interviewers memories enhanced too? Applied Cognitive Psychology, 8, 13-24.

Koehnken, G., Schimmossek, E., Aschermann, E., & Hofer, E. (1995). The cognitive interview and the assessment of the credibility of adults' statements. Journal of Applied Psychology. 80, 671-84.

Koriat, A. & Goldsmith, M. (1994). Memory in naturalistic and laboratory contexts: distinguishing accuracy oriented and quantity oriented approaches to memory assessment. Journal of Experimental Psychology, 123, 397-315.

Krass, J., Kinoshita, S., and McConkey, K.M. (1989) Hypnotic memory and confident reporting. Applied Cognitive Psychology, 3, 35-51.

Mantwill, M., Koehnken, G., & Aschermann, E. (1995). Effects of the cognitive interview on the recall of familiar and unfamiliar events. Journal of Applied Psychology, 80, 68-78

McCauley, M.R., & Fisher, R. (1995). Enhancing children's memory with the revised cognitive interview. Journal of Applied Psychology (4), 510-517

Memon, A. (1996). The cognitive interview. In O. Hargie. A handbook of communication skills. Routledge, London.

Memon, A., & Koehnken, G. (1992). Helping witnesses to remember more: The cognitive interview. Expert Evidence: The International Digest of Human Behaviour, Science & Law, 1, 39-48.

Memon, A. & Stevenage, S. (1996) Interviewing witnesses: What works and what doesn't? Psycholoquy 96.7.06.witness-memory.1.memon.

Memon, A. Cronin, Ó, Eaves, R. & Bull, R. (1996a) An empirical test of the mnemonic components of the Cognitive Interview. In G.M. Davies, S. Lloyd-Bostock, M. McMurran & C. Wilson (Eds.). Psychology and Law: Advances in Research. Berlin: De Gruyter.

Memon, A., Holley, A., Wark., Bull, R. & Koehnken, G. (1996b). Reducing suggestibility in witness interviews. Applied Cognitive Psychology, 10, in press

Memon, A., Wark, L., Holley, A., Bull, R. & Koehnken, G. (1996c). Interviewer behaviour in investigative interviews. Psychology, Crime and Law. 3, 181-201.

Memon, A, Cronin, Ó, Eaves, R., & Bull, R. (1993). The cognitive interview and child witnesses. In G.M. Stephenson & N.K. Clark (Eds.) Children, evidence and procedure. Issues in Criminological & Legal Psychology. No. 20. Leicester, UK: British Psychological Society.

Memon, A., Milne, R., Holley, A., Bull, R., & Koehnken, G. (1994). Towards understanding the effects of interviewer training in evaluating the cognitive interview. Applied Cognitive Psychology, 8, 641-659.

Memon, A. Wark, L., Bull, R., & Koehnken, G. (in press). Isolating

the effects of the cognitive interview. British Journal of Psychology

Milne, R., Bull, R., Koehnken, G., & Memon, A. (1995). The cognitive interview and suggestibility. In G.M. Stephenson & N.K. Clark (Eds.), Criminal behaviour: Perceptions, attributions and rationality. Division of Criminological & Legal Psychology Occasional Papers, No. 22. (pp. 21-27).Leicester, UK: British Psychological Society.

Norman, D., & Bobrow, D. (1978). Descriptions : An intermediate stage in memory retrieval. Cognitive Psychology, 11, 107-123.

Orne, M.T., Soskis, D.A., Dinges, D.F. and Orne,E.C. (1984). Hypnotically induced testimony. In G.L. Wells and E.F. Loftus (Eds), Eyewitness Testimony: Psychological Perspectives. pp. 171-215 New York: Cambridge University Press.

Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart, & Winston.

Richardson, S.A., Dohrenwood, B.S., & Klein, D. (1965). Interviewing: its forms and functions. New York: Basic Books.

Roethlisberger, F.J. and Dickson, W.J. (1946). Management and the worker. Cambridge: Harvard University Press.

Rogers, C.R. (1942). Counselling and Psychotherapy. Boston: Houghton-Mifflin

Tulving, E. (1974). Cue-dependent forgetting. American Scientist, 62, 74-82.

Tulving, E., & Thomson, D.M. (1973). Encoding specificity and retrieval processes in episodic memory. Psychological Review, 80, 353-370.

Appendix I: Interview Protocols

Each interviewer worked from a protocol which outlined the interview procedure to be followed. The interviewers were also instructed to take notes on the protocols. The information is summarised in the protocols.

Structured Interviews

Greet and Rapport -

Explain purpose-

Transfer Control -

Free Report -

Questioning phase - about what the interviewees said during their free report.


Cognitive Interviews

Essentially these followed the same procedure as the Structured Interviews with the following additions:

Report everything -encourage interviewees to report every detail they can remember, even partial information.

Reinstate Context - Before initiating a free report ask interviewees to think back to the original event and try and have an image of the video in their mind as they described it.

Repeat Report everything instruction ( prior to the questioning phase).

Reinstate Context with image probing - ask interviewees to reinstate context before the questioning phase. Used context reinstatement to probe specific images i.e. "Think about what the car looked like. Can you get a picture of the car in your mind?" before asking specific questions.

Untrained Interviewers

Essentially the untrained group of interviewers followed the same protocol as the Structured interviewers except that they were not instructed to transfer control or give the do not fabricate instruction.


This research was supported by a grant from the British Economic and Social Research Council (ESRC) R000234290. We are grateful to the staff and students of Taunton College, Hampshire. We thank Rebecca Milne for her contribution to the research, Debra Poole, Mona Mantwill and Ron Fisher for comments on earlier drafts and our interviewers for data collection. A portion of this paper was presented at the British Psychological Society Cognitive Section conference in Bristol, September, 1995.


[1] Address for correspondence.

[2] Fisher et al (1987a) document several characteristics of the "standard" police interview among Miami police officers such as constant interruptions when an eyewitness was giving an account, excessive use of question-answer format, and inappropriate sequencing of questions. Interestingly, George (1992) came up with a similar classification of the standard interview based on his analysis of interviews conducted by police officers in England.

[3] We used female interviewers due to practical constraints. We appreciate gender of interview may play an important role but this was not the focus of the current study.

[4] There were no significant differences between the three interviewers in any of the groups.

[5] All interviewers in the present study were asked to warn witnesses not to fabricate responses