In a forensic setting a witness is a key contributor. Much research has recently been directed to helping the witness achieve as full and accurate a recall as possible. One of the most promising techniques to emerge is the cognitive interview (CI). Tests of the cognitive interview with young adults suggest that it generates consistent and significant gains in the amount of correct information recalled. However, more recent studies (with adults and children) suggest that the gains are accompanied by an increase in errors and confabulations. These findings have important implications for the cognitive interview as a forensic tool. The following target article critically examines the evidence and raises theoretical and methodological issues arising from work on the CI. In light of the Recovered Memory debate, broader practical implications of this work are considered.
1. The role of witnesses is an important one. Their testimony can mean the difference between the police achieving a valid outcome or a culprit getting away with a crime (Wagenaar, 1988; Loftus & Ketchum, 1991). It can mean the conviction of a guilty person or the release of an innocent one. Witnesses are also vulnerable individuals, however, who can be influenced by poor retention of information, misleading information, and inappropriate forms of questioning (e.g., Gudjonsson, 1992; Ceci & Bruck, 1995). A large body of research has been conducted to provide interviewers with improved methods for eliciting information from their witnesses (see Goodman & Bottoms, 1993; Koehnken, 1995; Zaragoza, Graham, Hall, Hirschman & Ben-Porath, 1995 for reviews). Amongst these methods the cognitive interview (CI) is perhaps the most exciting development (Fisher & Geiselman, 1992; Memon & Koehnken, 1992). The CI is intended to facilitate recall, and initial tests show that it generates consistent gains in the amount of information that can be gathered from a witness. This target article examines the evidence for the CI and, in the light of more recent work, it attempts to identify conditions under which the procedure may be beneficial as well as the contexts in which it may be less helpful.
2. The CI was first developed in the USA by Geiselman and his colleagues (Geiselman, Fisher, Firstenberg, Hutton, Sullivan, Avetissian & Prosk, 1984; Fisher & Geiselman, 1992). It represented the alliance of two quite distinct fields of study. The concerns of social psychologists when managing a face-to-face interaction were integrated with what psychologists knew about the way we remember things. The social psychological concerns are embodied in what is referred to as a structured interview (SI). This consists of a phased procedure (free report, questioning and second retrieval phases) and incorporates techniques to facilitate communication. These include rapport building, which is designed to increase the transfer of control from the interviewer to the witness, and the use of a questioning strategy which is guided by the witness's own free report rather than being based on a predefined protocol.
3. The CI uses the same communicative components as the SI but in addition incorporates a number of cognitive strategies. These have several theoretical bases. First, it is known that a retrieval cue is effective to the extent that there is an overlap between it and the encoded information; with this in mind, the CI encourages the reinstatement of the original encoding context so as to increase the accessibility of stored information (Tulving & Thomson's Encoding Specificity Hypothesis, 1973). This is compatible with the "Headed Records" Model (Morton, Hammersley & Bekerian, 1985) in which "records" are discrete units of information containing event-relevant data which can be accessed via "headings". Headings are searched using "descriptions" (Norman & Bobrow, 1978) until a match is obtained. Information about context may form part of the description and may facilitate access to the original record. The reinstatement of context -- in which witnesses are encouraged to recreate the scene, report their feelings and the sequence of activities during an event -- is an attempt to use context cues to aid recall. Recent evidence suggests that the effectiveness of context reinstatement relies heavily upon the compatibility between affective states at the time of encoding and testing. In other words, how similar the environment feels may be more critical than how similar it looks (Eich, 1995) although there is some dispute about this (Smith, 1995).
4. A somewhat different perspective upon which the CI relies is Multiple Trace Theory (Bower, 1967). This suggests that rather than consisting of discrete and unconnected incidents, our memories are made up of a network of associations; consequently, there are several ways in which a memory could be cued (such models rely on the concept of activation spreading throughout a network). On the basis of this, the CI incorporates multiple retrieval strategies to exploit the many cues and thereby aid recall. One such strategy is to recall an event from different starting points -- from the beginning, middle and end -- thus working backward and forward in time. This change in temporal perspective may allow witnesses to recall information which had been overlooked or judged unimportant at a previous recall attempt. However, as pointed out by Bekerian and Dennett (1993), altering the temporal order of recall works in opposition to the strategy of context reinstatement in which details are recreated so as to be as close to their true order and appearance as possible. This may account for the finding that reverse order recall appears relatively ineffective (in terms of yielding additional information) when used in conjunction with the context reinstatement technique (Memon, Wark, Bull & Koehnken, in press a).
5. Before discussing the final theoretical model that has influenced the CI, we need to consider the implications of repeatedly asking a witness to recall an event. Laboratory studies of memory have established two main effects of repeated testing. The first is "reminiscence", or the recall of material that was not retrieved in an earlier attempt. When this new information exceeds the amount of information that is forgotten, the "hypermnesia" effect has occurred (Payne, 1987). Both these phenomena have been demonstrated in eyewitness contexts using video-taped (Scrivner & Safer, 1988) and staged criminal scenarios (Turtle & Yuille, 1994). It is possible that hypermnesia and reminiscence in the CI situation is the result of using different mnemonic strategies or multiple retrieval opportunities throughout the interview (see Payne, 1987). The overlap in theoretical explanations for CI effects is clearly apparent.
6. The last model to influence the CI is Schema Theory (Schank & Abelson, 1977) according to which familiar events have a schema or script (based on prior experience) and this script guides the encoding of an event by organising information into a hierarchy of slots. These slots prime the witness as to the details which should be encoded, but more than this, a witness's prior expectations allow him to fill in empty slots with default information. In a similar way, the schema can guide retrieval of information by providing an organised information system to search. Multiple retrieval techniques, as well as techniques to alter the temporal order of recall and to probe mental images, can all be seen as attempts to activate this information system at different points in order to maximise recall. In this respect, Schema Theory is similar to the Multiple Trace theory discussed in section IV. Where the two theories differ, however, is in the concept of default information. Recall, according to Schema Theory, is prone to the inclusion of information that is assumed on the basis of prior expectations. In many instances this schema-consistent default information may be correct. However, on occasion, schema-consistent information will be recalled in place of unlikely, surprising or schema-inconsistent actual details. The schema theory thus supplies us with a way of explaining errors (incorrect details) and confabulations (inserted details) in recall.
7. The results of early work on the CI have led to the development of an enhanced version of the CI (Fisher & Geiselman, 1992; Fisher, Geiselman & Amador, 1989; Fisher, Geiselman, Raymond, Jurkevich & Warhaftig, 1987); it is this enhanced version that is the recommended procedure. This differs from the original CI in various respects. For example, communicative techniques are incorporated which transfer the control of the interview to the witness. In addition, interviewers allow the witness to dictate the pace of the interview by using open-ended questions, not interrupting replies, timing questions and comments carefully, and using the witness's recall as a guide for subsequent questioning. On the basis of Paivio's (1971) work on the effectiveness of mental imagery in recall, the enhanced CI also encourages witnesses to create an image of the scene during questioning. Interviewers are directed to probe the image and elicit detailed information (see Fisher & Geiselman, 1992, Memon, in press a for examples). The CI therefore focuses primarily on the use of explicit memory (i.e., strategies that facilitate the conscious and intentional recollection of an event) although this does not rule out implicit processes.
8. The effectiveness of the CI has been extensively evaluated in Germany as well as in the USA and Britain. Earlier studies reported positive and significant effects both in and out of the laboratory. With relatively little training, cognitive interviewers obtained up to 35% more correct information from adult witnesses than untrained interviewers in the recall of simulated events witnessed days earlier. Perhaps more important, the CI showed no increase in the number of errors obtained (Geiselman, Fisher, MacKinnon & Holland, 1985; 1986). With the enhanced CI the results were even more impressive, yielding 45% more correct information compared to the standard CI (Fisher et al., 1987).
9. 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). Researchers working in both the laboratory and the field have attempted to replicate these findings with varying degrees of success (George & Clifford, 1991; Mantwill, Koehnken & Aschermann, 1995; Memon, Cronin, Eaves & Bull, 1996; Gwyer, Clifford & Dritschel, 1995). Limitations of the CI first became apparent when it was applied to child witnesses where it was consistently found to increase the incidence of errors or confabulated details (Koehnken, Finger, Nitschke, Hofer & Aschermann, 1992; Memon et al., in press a; Milne, Bull, Koehnken & Memon, 1995; Fisher & McCauley, 1995; McCauley & Fisher, 1995). The only study of the enhanced CI reported with child witnesses where there was no such increase in errors was one conducted by Saywitz, Geiselman and Bornstein (1992). In the latter study, the CI (original version) was found to benefit older children (11-12 year olds) more than younger children (7-8 year olds). In our own studies with children (e.g., Memon et al., in press a, c) we found that children aged 8-9 years are able to make effective use of the CI techniques but at the cost of a small increase in errors. With child witnesses, one of the keys to a successful interview (CI or SI) appears to be effective rapport building and communication. With this in mind some researchers recommend a "practice interview" prior to the main interview (see Saywitz et al., 1992; McCauley & Fisher, 1995). Others emphasise the importance of interviewer training and practice in the use of appropriate rapport building techniques (Memon et al., in press a). It is clear that some special considerations are required when using the CI with child witnesses.
10. So far we have suggested that when weighed against the previous evidence, recent research suggests that the advantage gained from the CI may be outweighed by an increase in errors or confabulated details. The remainder of this paper examines the factors that determine the effectiveness of the CI in five main categories: standards of comparison, interviewer variables, interview technique, questioning procedures, and errors.
11. The CI generates important increases in the quantity of correctly recalled information about an event. However, there are methodological questions about the standard or control against which the CI performance is evaluated. As suggested earlier, studies show a CI advantage when the reports obtained by CI interviewers are compared to those obtained by untrained interviewers. It is noteworthy that this pattern is consistent whether those untrained interviewers are complete novices (student interviewers, Geiselman et al., 1984) or they have some experience in an interview situation (police interviewers, Geiselman et al., 1985). These results imply that the CI techniques elicit a better and fuller recall than an untrained baseline. The finding that police interviewers untrained in the CI technique are no better than novice interviewers is not intended to imply that their experience is worthless; indeed, police officers are very sensitive about this (Memon, Bull & Smith, 1995). It is likely, however, that the positive effects of their experience are offset by the negative effects of poor interview techniques (Memon, Milne, Holley, Bull & Koehnken, 1994).
12. The CI findings suggest that the untrained interviewers are doing a rather poor job. This in turn suggests that this untrained group may not offer an adequate comparison with the CI. Indeed, for professional practitioners to be confident in adopting the CI technique, they need to know whether it produces gains relative to the techniques they currently use. Recent studies have compared the amount of information retrieved using CI techniques with the amount retrieved by structured interviewers (SI) who are trained in all of the CI techniques except the cognitive mnemonic strategies. That is, both CI and SI interviewers use the communicative techniques of rapport building and transfer of control, but in addition, the CI interviewers make use of the cognitive retrieval techniques of context reinstatement, reporting in detail, multiple retrieval and imagery. The use of SI interviewers as a comparison group helps us to determine whether the CI gains noted previously can be attributed to the cognitive retrieval techniques alone. It is interesting to note that when the comparison group receives a structured interview rather than a standard interview, the CI advantage is reduced.
13. Concerning the issue of what constitutes an appropriate control group, three factors are likely to influence effectiveness: the interviewer's (i) the actual knowledge, (ii) perceived knowledge, and (iii) perceived status.
14. In an applied setting, interviewers having some knowledge of the event can be seen as positive. They may be able to provide prompts to aid the witness and their knowledge may allow them to corroborate the witness's story. Unfortunately, there are negative consequences of interviewer knowledge as well. Being human, interviewers are undoubtedly influenced by their beliefs (as documented by Rice, 1929) and may mislead the witness by providing false prompts (Ceci, Leichtman & White, in press, report some preliminary data which support this). Moreover, the interview is likely to consist of more closed questions designed to get facts rather than open questions designed to elicit a narrative. This can alter the structure of the cognitive interview with negative consequences (e.g., Memon et al., in press a).
15. A second factor linked to the interviewer's knowledge of the event depends on the witness's perception of this knowledge, and this can be influenced by the interviewer's status (see Ceci, Ross & Toglia, 1987). If interviewers present themselves as a naive third party then witnesses are likely to feel more inclined to volunteer information, even though they may not be certain of its accuracy. Recent research has taken note of this by introducing the "transfer of control" technique into the enhanced CI procedure. This technique aims to increase the witness's perceived status in the interview situation and may contribute to the CI advantage. In Memon et al.'s studies both the CI and SI interviewers use this technique. Hence, the way interviewers present themselves can set up demand characteristics in the interview situation which influence the demonstration of the CI advantage. This hypothesis should be explored in further research.
16. The final aspect of the cognitive interview that needs to be looked at is the nature, amount and timing of questioning in the interview. Traditionally, the questioning phase follows a free recall phase. The interviewer makes use of open-ended questions and uses information from the free recall as prompts to aid the witness. One concern is the extent to which interviewers make use of these questioning techniques and, specifically, how training affects the number of questions an interviewer asks. Evidence on this point is remarkably mixed. Early studies indicate that their CI interviewers actually ask fewer questions than the untrained control group. These untrained interviewers characteristically make use of a number of poor interview techniques in combination with a large number of questions (Fisher, Geiselman & Raymond, 1987). More recent research with a more carefully controlled comparison group suggests that interviewers who have received training of any sort (cognitive or structured) ask more questions than untrained interviewers (Memon, Wark, Bull & Koehnken, 1995).
17. In addition to the number of questions asked in different interview conditions, it is important to look at the effect of interview type across questioning sessions. As documented in the 1992 Government recommended Memorandum of Good Practice (England and Wales), the cognitive interview uses a variety of techniques in sequence, proceeding from free recall to open questions to closed questions. This allows the witness a number of attempts to recall information, and this may lead to the demonstration of the well-established effect of reminiscence, in which new details may be elicited with each successive recall attempt (Payne, 1987). The extent to which this explanation accounts for the CI effects, however, has not been demonstrated until recently. Memon et al. (1995) compared the performance of CI, SI and untrained groups across two discrete sessions (questioning and second retrieval attempt). When data from the second retrieval attempt were removed from the analyses, there was an overall advantage of the CI and SI interview relative to the untrained group. However, when control subjects were given the opportunity to engage in a second retrieval attempt, their performance improved to the level of the cognitive group. Indeed, several studies have suggested that the CI is only effective at the first questioning phase (Memon et al., in press a; Milne et al., 1995). This suggests that the CI advantage is related to the number of retrieval attempts. Interestingly, when it comes to repeated testing (across different interview sessions), the CI is only beneficial the first time it is used. It does not lead to additional gains in information on a second interview as compared to a structured interview (Memon et al., in press a; McCauley & Fisher, 1995). The effects of repeated testing and multiple retrieval are clearly important from the theoretical and forensic perspective and warrant further investigation.
18. The previous section raises the question of what actually underlies the CI advantage documented to date. What is it about the technique that gives rise to the gains seen in the earlier studies? Previous researchers have been unable to identify a discrete component of the CI responsible for the gains in recall (Geiselman et al., 1986; Memon et al., in press b). However, some research has indicated the importance of context reinstatement when used in isolation from the other CI cognitive techniques (see Davies & Thomson, 1988 for a review). Surprisingly, however, police officers seldom use the context reinstatement instruction when conducting a cognitive interview (Memon et al., 1994; Turtle, 1995), at least not in the way it is defined in research studies (but see George, 1992). Further research is accordingly required before any conclusions can be drawn as to the relative importance of the cognitive techniques comprising the CI.
19. More recent findings, however, question the extent to which the CI advantage is due to the use of cognitive techniques alone (Memon et al., 1995; Fisher & McCauley, 1995). For example, in the Memon et al. study the performance of CI interviewers was compared to that of SI interviewers, with both groups receiving extensive training over two days, including experiential role playing and feedback from video assessment. The results suggest that there is no difference between the two groups of adult witnesses (college students) in terms of overall amount recalled, accuracy of recall, number of errors (falsely recalled details) or number of confabulations (made-up details). Moreover, Memon et al. (1995) found that both CI and SI groups elicited more information overall, and more correct information relative to an untrained group. This was offset, however, by both CIs and SIs producing a significantly higher number of errors and confabulations than an untrained group. This study raises the issue of whether the CI advantage could be attributed to the cognitive retrieval techniques at all or whether it may be the result of a facilitation of communication. Alternatively, are the gains a result of improved memory and improved communication, as Fisher and McCauley (1995) suggest?
20. Memon et al.'s findings suggest that the CI advantage documented in the literature may not be entirely a result of cognitive mnemonics but may rest in part on the establishment of an effective rapport between interviewer and witness. At this point, however, it is important to look at the work of Koehnken and colleagues (Koehnken, & Zorberbier, 1994; Koehnken, Schimmossek, Aschermann & Hofer, in press; Mantwill et al., 1995). These investigators have also compared the performance of a CI group with an SI group, the latter trained in rapport-building, the use of different questioning strategies, and organising interviews. Unlike Memon et al. (1995), Koehnken et al. do find significant differences in favour of the CI group (as do Milne et al., 1995; and Memon et al., in press a, with child witnesses). This indicates an advantage when cognitive strategies are used in conjunction with communicative techniques rather than when the same communicative techniques are used alone. The reasons for the differences in results are unclear; a close look at the methodologies, however, suggests that we need to be cautious about making direct comparisons between these studies. For example, from a description of the interviewer training for each study, it is not clear whether the CI and SI groups differ in the same way across studies. There may be other factors that vary. It is possible, for example, that SI interviewees make implicit use of the CI techniques in some cases. Indeed, Cohen and Java (1995) and Memon, Wark, Holley, Bull, and Koehnken (in press b) provide evidence of the spontaneous use of context reinstatement by witnesses during the free recall phase of an interview. The evidence surrounding this issue is clearly contradictory; consequently, the suggestions that the CI advantage amounts to a case of facilitated communication remain speculative. However, the theoretical and practical implications of such a suggestion have great importance. One way of examining the value of facilitated communication is to hold it constant across all interview groups by comparing the CI and SI with a rapport-only control group; we encourage replications of the studies described here to help clarify the mechanisms underlying the CI effects so far.
21. Together with the cognitive and the communicational components, a third cluster of components may contribute to the CI effects. These components can be viewed as motivational, affecting both the interviewer and the witness. Interview training of any kind results in an increase in motivation to be a successful interviewer and to demonstrate that the training has worked (Rosenthal & Rubin, 1978). This can make trained interviewers ask more questions, probe for information and create demand characteristics -- a social pressure to give a desired response -- that induce the witness to say more (Cronin, Memon, Eaves & Bull, 1992). This can have negative effects on error rates (see section VII).
22. The witness is also clearly subject to motivational effects. For example, if the event is particularly distressing (e.g., Terr, 1990), the witness may not wish to recreate the event in the way that is required in a CI (c.f., Christianson & Hubinette, 1993; Yuille & Tollestrup, 1992). Fisher and Geiselman (1992) have pointed out that the CI should only be used with co-operative witnesses. However, it is not always clear whether or not someone wishes, and is able, to remember an event (see Loftus & Ketchum, 1991; Lindsay & Read, 1994; Kihlstrom, in press). It is likely in such cases, that an interviewer (or therapist) will use some leading (and possibly misleading questions) as prompts or cues. This in turn is likely to increase the amount of correct information recalled, but with an accompanying increase in the number of errors. We would therefore caution against using the CI for recovering memories of trauma until its utility has been systematically investigated. In the recovered memory context, the choice of interview technique should be determined by the aims of the interview. If the purpose is to elicit details which are accurate but cannot be corroborated, then perhaps the version of the CI described in the current paper is not suitable (see Pennebaker & Memon, in press, for a discussion of this issue).
23. In summary, whereas previous research focused on the facilitative effects of the cognitive strategies of the CI, current research is placing more emphasis on the contribution of social elements of the CI which appear in some way to facilitate the use of cognitive strategies. Further research is clearly needed to identify (a) the conditions under which an advantage of the CI may be present and (b) the control techniques which generate performance comparable to the CI.
24. Studies of the CI have used different dependent variables. Early work relied on calculating the number of correct and incorrect details recalled (Geiselman et al., 1984; Geiselman et al., 1986) and found that, relative to an untrained group, the CI elicits more correct items and fewer incorrect ones. Later studies motivated by the use of the CI in a forensic setting have shown that one must distinguish between errors, when a real detail is reported incorrectly (i.e., a blue car is described as black) and a confabulation in which a detail is inserted from imagination (Koehnken et al., 1994; Memon et al., in press a; Gwyer et al., 1995; see also, Gudjonsson & Clare, in press). The difference between an error and a confabulation rests on both a theoretical and a forensic distinction. Theoretically, an error involves the misremembering of encoded information whereas a confabulation involves the report of a fact which is created rather than being drawn from the encoded information. Forensically, the difference between an error and a confabulation can be illustrated by using the example that an error involves the inaccurate recall of an event that actually happened (i.e., reporting that a man pulled the trigger when the culprit was female) whereas a confabulation involves inserting a fictitious event into the recall. This distinction becomes relevant when an entirely false narrative is created with the use of some misleading cues (see Ceci et al., 1994 for an example).
25. Studies have also included a measure of report accuracy or reliability (proportion of correct information relative to total information) to complement the measure of the absolute amount of correct information. Use of these various measures creates quite a complicated picture. Most report an increase in the absolute amount of correct information with the CI. However, these studies have found that the CI generates more correct information at the cost of generating more errors/confabulated details (Mantwill, Koehnken & Aschermann, 1995; Koehnken et al., in press; Gwyer et al., 1995; Memon et al., in press a; Bekerian, Dennett, Reeder, Slopper, Saunders & Evans, 1994). If we look at the absolute numbers of errors compared to correct details in these studies, it is clear that the CI produced a greater increase in correct details than in incorrect details. In the Memon et al. study, we felt that a plausible explanation for the increase in errors was that the CI technique shifted the criterion for responding by influencing confidence and willingness to report information (the theoretical basis for this is illustrated by signal detection theory as noted by Bekerian & Dennett, 1993). Interestingly, a similar effect (an increase in correct and incorrect details) 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. The same may be the case with the CI, since the "report everything" instruction specifically asks witnesses to report all details.
26. The effect of CI on the response criterion requires further investigation. It is possible that the errors are a result of demand characteristics or social pressure to give a desired response (Cronin et al., 1992), but it is not yet clear which aspects of the CI procedure are responsible for the increases in errors. A number of possibilities are suggested by the work of Memon and co-workers. One possibility is that the errors are due to more detailed probing by CI interviewers. Asking detailed questions typically increases errors (e.g., Davies, Tarrant & Flin, 1989). At the same time, it could be argued that responses to specific questions are less likely to be associated with confidence as witnesses do not have the opportunity to "filter out" their responses as they do in free recall situation (see Koriat & Goldsmith, 1994). The errors may also be related to the type of event being recalled. It is possible that events for which children have pre-existing knowledge structures or schemata (such as the magic show used in the Memon et al. in press a study) may be especially subject to errors with a CI). Mantwill et al. (1995) have specifically addressed this question by looking at the effects of event familiarity on recall with a CI but they found no effects of this variable on error rates (although there were significantly more errors in the CI group overall). However, this does not rule out the possibility that errors may be due to recall of schema-consistent information. The extent to which prior knowledge may induce errors or confabulated details in an interviewee's account has recently been considered in relation to false or implanted memories (Pezdek, 1995).
27. Another plausible hypothesis for the increase in error rate with the CI is that the errors may be linked to the use of imagery techniques. The use of such techniques in isolation has been shown to increase errors (Bekerian, Dennett, Hill & Hitchcock, 1992); linked with the use of misleading questions, imagery may increase source confusions (Carris, Zaragoza & Lane, 1992; Ceci, Loftus, Leichtman & Bruck, 1994). Again, this finding has important implications in light of concerns about the accuracy of memories recovered with techniques such as guided imagery (Lindsay & Read, 1994).
28. Note that in the Memon et al. (in press a) study it was not possible to separate the use of imagery instructions from the use of specific question probes because the two were related. We were, however, able to locate the nature of the errors that were being made. The children made more errors about persons than any other type of information, an effect that has been replicated in several recent studies (e.g., Milne et al., 1995). Again, previous literature suggests that children (Davies et al., 1989; Gee & Pipe, 1995) and adults (MacLeod, Frowley & Shepherd, 1994) have difficulty in describing persons.
29. How should the increase in errors be interpreted by practitioners? First of all, it should be noted that the CI tends to increase errors rather than confabulated details. An increase in confabulated details (e.g., saying the suspect had an accomplice when there was none) may have quite different ramifications in a forensic context than an error (e.g., describing a purple jacket as a blue one). Second, if we look at the absolute number of correct details compared to errors in the various studies, the gains appear to outweigh the risks (e.g., in the Memon et al., in press a study, there was only one error for every six correct details). Thirdly, the majority of studies find no difference in accuracy rates across cognitive and structured interview conditions. In other words, the proportion of correctly or incorrectly recalled details does not seem to vary between CI and SI. While procedures such as the CI aim to increase the amount of information reported, in a forensic context it is the accuracy of the information that is crucial (Koriat & Goldsmith, 1994). In a recent review, Koriat and Goldsmith (1996) demonstrate how critical it is to separate the measure of quantity of information, which fits the "storehouse metaphor" of traditional research, from applied research on everyday memory processes. In the legal context, the accuracy with which a detail is recalled is paramount because the interviewer cannot tell what is true and what is not (see Ceci & Bruck, 1995 for a discussion of this).
30. We suggest that practitioners carefully assess the purpose of interviewing prior to using a cognitive interview. If the aim is simply to obtain as much detail as possible for the purposes of eliciting leads in an investigation, then the CI is likely to be useful. However, where the accuracy of small details is critical perhaps a more cautious approach should be adopted. Whenever possible, researchers should attempt to corroborate details, particularly when they pertain to descriptions of persons (several studies have found errors with this type of information in a CI, see Milne & Bull, 1995; Milne et al., 1995). Regarding young witnesses, we suggest that the CI can be used with children over the age of 8 years (Memon et al., in press a; Saywitz et al., 1992), but some caution should be exercised when interviewing younger children (see Memon et al., 1993).
31. In making recommendations it is necessary to distinguish the needs of the researcher from those of the practitioner. Further tests of the CI are needed if researchers are to advise practitioners in the use of this procedure. Research should be guided by the need to evaluate the effectiveness of the CI technique in the context for which it is created rather than in an artificial laboratory situation. At the same time, tighter controls are needed in order to understand the conditions in which the techniques may be effective. The impact of the CI following a delay (longer than a few days) between the event and test, and multiple interview sessions is not clear. In addition, studies need to examine the potentially interesting issue of interviewer differences. Finally, it is important for researchers to draw the attention of practitioners to null and negative effects such as increases in errors (see Herrmann, 1995 for a general discussion as to the importance of this) and initiate studies designed to provide an explanation for these effects.
32. It is also important to appreciate the broader implications of work on the CI. We have already discussed the concerns surrounding both standard interview practices for witnesses and therapeutic interview practises in light of the recovered memory issue. Clearly, both situations call for a valid, reliable and appropriate interviewing procedure. Without doubt, trained CI interviewers can encourage the interviewee to report more correct information, and indeed to say more overall. However, the current review brings out the limitations of the technique. It is inevitable that a technique which produces significant gains in information will produce an increase in errors and intrusions and we urge interviewers to take measures to minimise the frequency and impact of these details. Clearly, further work is required to identify the conditions in which the CI is of maximum benefit.
NOTE: The first author is currently conducting research in collaboration with Steve Ceci and his team on the utility of CI in interviewing preschoolers.
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., Hill, K., & Hitchcock, R. (1992) Effects of detailed imagery on simulated witness recall. In F. Losel, D. Bender, & T. Bleinsener (Eds.). Psychology and Law: International Perspectives. Berlin: Walter de Gruyter.
Bekerian, D.A., Dennett, J.L., Reeder, C., Slopper, K., Saunders, H., & Evans, L. (1994) The influence of the cognitive interview technique on attributions of memory. Paper presented at the Annual Conference of British Psychological Society, April, Brighton.
Bower, G. (1967) A multicomponent theory of memory trace. In K. W. Spence and J.T. Spence (Eds.) The Psychology of Learning and Motivation, Vol 1, New York: Academic Press.
Carris, M., Zaragoza, M., & Lane, S. (1992) The role of visual imagery in source misattribution errors. Paper presented at the annual meeting of the Midwestern Psychological Society, Chicago, IL.
Ceci, S.J., & Bruck, M. (1995) Jeopardy in the Courtroom. New York: American Psychological Association books.
Ceci, S., Loftus, E., Leichtman, M. & Bruck, M. (1994) The possible role of source misattributions in the creation of false beliefs among preschoolers. International Journal of Clinical and Experimental Hypnosis, 42, 304-20.
Ceci, S., Leichtman, M. & White, T. (in press) Interviewing preschoolers. In Peters, D. (Ed.), The Child Witness in Cognitive, Social and Legal Context. Netherlands, Kluwer.
Ceci, S.J., Ross, D.F., & Toglia, M.P. (1987) Age differences in suggestibility: Narrowing the uncertainties. In S.J. Ceci, M.P. Toglia, & D.F. Ross (Eds.), Children's Eyewitness Testimony, New York: Springer.
Christianson, S. A. & Hubinette, B. (1993) Hands up! A study of witnesses' emotional reactions and memories associated with bank robberies. Applied Cognitive Psychology, 7, 365-379.
Cohen, G., & Java, R. (1995) Memory for medical history: Accuracy of recall. Applied Cognitive Psychology, 9, 273-288.
Cronin, O., Memon, A., Eaves, R., & Bull, R. (1992, September) The cognitive interview with child witnesses: a child centred approach? Paper presented at the Third European Conference of Psychology and Law, Oxford University, Oxford.
Davies, G., Tarrant, A., & Flin, R. (1989) Close encounters of the witness kind: Children's memory for a simulated health inspection. British Journal of Psychology, 80, 415-529.
Davies, G.M., & Thomson, D.M. (Eds). (1988) Memory in Context: Context in Memory, Chichester, John Wiley and Sons Ltd.
Eich, E. (1995) Mood as a mediator of place dependent memory. Journal of Experimental Psychology: General, 124, 293-308.
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(5), 722-727.
Fisher, R.P., Geiselman, R.E., & Raymond, D.S. (1987) 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. (1987) Enhancing eyewitness memory: refining the cognitive interview. Journal of Police Science and Administration, 15, 291-297.
Fisher, R.P. & 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.
Gee, S., & Pipe, M.E. (1995) Helping children to remember: The influence of object cues on children's accounts of a real event. Developmental Psychology, 31(5), 746-758.
Geiselman, R.E., Fisher, R.P., Firstenberg, I., Hutton, L.A., Sullivan, S.J., Avetissian, I.V. & Prosk, A.L. (1984) Enhancement of eyewitness memory: an empirical evaluation of the cognitive interview, Journal of Police Science and Administration, 12, 74-79.
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.
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.
George, R.C. (1992) A field evaluation of the cognitive interview. Unpublished Master's Thesis, University of East London.
George, R.C., & Clifford, B. (1991) A field comparison of the cognitive interview and conversation management. Paper presented at the British Psychological Society Annual Conference, Bournemouth, U.K.
Goodman, G. & Bottoms, B. (1993) Child victims, child witnesses: Understanding and Improving testimony. New York: The Guilford Press.
Gudjonsson, G. (1992) The psychology of interrogations, confessions and testimony. Chichester, Wiley.
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.
Herrmann, D. (1995) Doing good science and appreciating the basic applied relationship. Paper presented at the 1st biennial meeting of the Society for Applied Research in Memory and Cognition, University of British Columbia.
Kihlstrom, J. F. (in press) The Trauma Memory argument. Consciousness and Cognition.
Koehnken, G., Finger, M., Nitschke, N., Hofer, E. & Aschermann, E. (1992) Does a cognitive interview interfere with a subsequent Statement Validity Analysis? Paper presented at the American Psychology and Law Society meeting, San Diego.
Koehnken, G. (1995) Interviewing adults. In R. Bull & D. Carson (Eds.) Handbook of psychology in legal contexts. Chichester: Wiley.
Koehnken, G., Schimmossek, E., Aschermann, E. & Hofer, E. (in press) The cognitive interview and the assessment of the credibility of adults' statements. Journal of Applied Psychology.
Koehnken, G., Thurer, C., & Zorberbier, D. (1994) The cognitive interview: Are interviewers memories enhanced too? Applied Cognitive Psychology, 8, 13-24.
Koriat, A. & Goldsmith, M. (1996) Memory metaphors and the laboratory/real-life controversy: Correspondence versus storehouse views of memory. Behavioral and Brain Sciences, 19.2 (June, 1996).
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., & McConkey, K.M. (1989) Hypnotic memory and confident reporting. Applied Cognitive Psychology, 3, 35-51.
Lindsay, D.S. & Read, J.D. (1994) Psychotherapy and memories of child sexual abuse: A cognitive perspective. Applied Cognitive Psychology. 8, 281-338.
Loftus, E. & Ketchum, K. (1991) Witness for the defence. New York: St Martins Press.
MacLeod, M., Frowley, J., & Shepherd, J. (1994) Whole body perception and identification. In D. Ross, R.D. Read, & M. Toglia (Eds.). Adult eyewitness testimony: Current trends and developments. Cambridge, England: Cambridge University Press.
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.P. (1995) Facilitating children's eyewitness recall with the revised cognitive interview. Journal of Applied Psychology, 80, (4), 510-517.
Memon, A. (in press) The cognitive interview. In O. Hargie. A handbook of communication skills. London: Routledge. http://www.soton.ac.uk/~am/memon-in-press.html
Memon, A., Bull, R. & Smith, M. (1995) Improving the quality of the police interview: can training in the use cognitive techniques help? Policing and Society, 5, 53-68.
Memon, A., Cronin, O., Eaves, R., & Bull, R. (1996) An empirical test of the mnemonic components of the Cognitive Interview. In G.M. Davies, S. Lloyd-Bostock, M. McMurran & C. Wilson (Eds) Psychology, Law & Criminal Justice. (pp. 135-145). Berlin: De Gruyter. http://www.soton.ac.uk/~am/mceb1996.html
Memon, A., Cronin, O., 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 Crimological and Legal Psychology. No. 20. British Psychological Soeicty, Leicester, UK.
Memon, A., & Koehnken, G. (1992) Helping the witness to remember more: The cognitive Interview. Expert Evidence: the International Digest of Human Behaviour, Science and Law 1, 39-48.
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 a) Isolating the effects of the cognitive interview. British Journal of Psychology. http://www.soton.ac.uk/~am/mwbk-in-press.html
Memon, A., Wark, L., Holley, A., Koehnken, G. & Bull, R. (in press b) Context effects and event memory: How powerful are the effects: In D. Payne & F. Conrad (Eds.). Intersections in Basic and Applied Memory Research. Lawrence Erlbaum Associates. http://www.soton.ac.uk/~am/context-in-press.html
Memon, A., Wark, L., Holley, A., Koehnken, G. & Bull, R. (in press c) Reducing suggestibility in witness interviews. Applied Cognitive Psychology. http://www.soton.ac.uk/~am/mwhbk1995.html
Milne, R., & Bull, R. (1995, September) Interviewing children with mild learning difficulty with the cognitive interview. Paper presented at the Division of Criminological and Legal Psychology Annual Conference, Rugby.
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 and Legal Psychology, Occasional Papers, No. 22. British Psychological Society, Leicester, UK.
Morton, J., Hammersley, R.H., & Bekerian, D.A. (1985) Headed records: A model of memory and its failures. Cognition, 20, 1-23.
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., & Orne, E.C. (1984) Hypnotically induced testimony. In G.L. Wells, & 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 and Winston.
Payne, D.G. (1987) Hypermnesia and reminiscence in recall: A historical and empirical review. Psychological Bulletin, 101, 5-27.
Pennebaker, J.W., & Memon, A. (in press) Recovered memories in context: Thoughts and elaborations on Bowers and Farvolden. Psychological Bulletin. http://www.soton.ac.uk/~am/sub1995.html
Pezdek, K. (1995) What types of childhood event are not likely to be suggestively planted? Paper presented at the Annual meeting of the Psychonomic Society. Los Angeles.
Rice, S.A. (1929) Interviewer bias as a contagion. American Journal of Sociology, 35, 421-423.
Rosenthal, R., & Rubin, D.B. (1978) Interpersonal expectancy effects: The first 345 studies. Behavioural and Brain Sciences, 3, 377-415.
Saywitz, K.J., Geiselman, R.E. & Bornstein, G.K. (1992) Effects of cognitive interviewing and practice on children's recall performance. Journal of Applied Psychology, 77, 744-756.
Schank & Abelson, (1977) Scripts, plans, goals and understanding. Hillsdale, NJ: Erlbaum.
Scrivner, E., & Safer, M.A. (1988) Eyewitnesses show hypermnesia for details about a violent event. Journal of Applied Psychology, 73, 371-377.
Smith, S. (1995) Mood is a component of mental context: Comment on Eich. Journal of Experimental Psychology: General. 124(3), 309-310.
Terr, L. (1990) Too scared to cry: How trauma affects children and ultimately us all. New York: Basic Books.
Tulving, E., & Thomson, D.M. (1973) Encoding specificity and retrieval processes in episodic memory. Psychological Review, 80, 353-370.
Turtle, J. (1995, July) Officers: What do they want? What have we got? Paper presented at the 1st biennial meeting of the Society for Applied Research in Memory and Cognition, University of British Columbia.
Turtle, J., & Yuille, J. (1994) Lost but not forgotten details. Repeated eyewitness recall leads to reminiscence but not hypermnesia. Journal of Applied Psychology, 79, 260-271.
Wagenaar, W.A. (1988) Identifying Ivan: A case study in legal psychology. Hertfordshire England: Harvester Wheatsheaf.
Yuille, J.C. & Tollestrup, P. (1992) A model of the diverse effects of emotion on eyewitness memory. In S-A. Christianson (Ed.), The handbook of emotion and memory: Research and theory. Hillsdale, NJ: Erlbaum.
Zaragoza, M., Graham, J.R., Hall, G.C.N., Hirschman, R. & Ben-Porath, Y.S. (1995) (Eds.), Memory and testimony in the child witness Thousand Oaks, CA: Sage.