Fisher's (1996) commentary on the target article by Memon and Stevenage (1996) has raised two very important issues. The first concerns the suitability of the structured interview (SI) as a comparison group for the assessment of the cognitive interview (CI). The second issue concerns the relative importance of the absolute number of errors versus the error rate when discussing the accuracy of information elicited using the CI. We would like to take this opportunity to clarify our position on these two points.
1. In his commentary on our target article, Fisher (1996) discusses several issues which have been raised in previous commentaries. These concern the concepts of appropriate comparison and assessment of accuracy when using the cognitive interview (CI). These are such fundamental issues that we feel it important to try to clarify our thoughts further on these two points.
2. In order to be able to evaluate the effectiveness of a tool, we obviously need a standard against which it can be assessed. The CI has been assessed relative to two comparisons: the standard interview (currently widely used by interviewers) and the structured interview (SI -- based on the Memorandum of Good Practice and equivalent to the CI minus its cognitive mnemonic components). Fisher argues that the standard interview alone represents an appropriate comparison group for assessment of the CI. He suggests that the structured interview group fails as a meaningful comparison group on two counts: its comparison is neither driven by the use of the structured interview as a practical and valid alternative procedure, nor by any theoretically meaningful rationale.
3. Several issues concern us with Fisher's argument. First, it is by no means clear that the standard interview is as reliable a comparison as it would first appear. Our experience suggests that there are large individual differences in the quality of interviews among professional interviewers. Some interviewers are very skilled at eliciting information even without extensive training, while others are consistently poor (see Bull & Cherryman, 1996). The term "standard interview" may thus refer to interviews which range in quality, and consequently, we may sometimes be comparing CI against a good standard interview and other times against a poor one (see Kebbell & Wagstaff, 1996). Relative to this, the structured interview benefits from some consistency of style. Interviewers are trained in the techniques that they use and moreover, it is possible to control for the effects of training, interviewer motivation, and so on. While individual differences between interviewers undoubtedly remain, it is possible to look at these systematically at the analysis stage (see Memon, Wark, Holley, Bull & Koehnken, 1996b). In light of this, it may be argued that the SI actually provides a more uniform comparison group against which the CI may be assessed.
4. Fisher suggests that the SI fails as a practical and meaningful comparison for the CI because the SI is not currently used in the field. Such a comparison could not therefore tell us whether the CI was any better than an existing interviewing technique (Fisher, par. 3). In fact, we would like to draw the reader's attention to the extensive use of the SI both in North America and in Britain. For instance, the SI resembles the techniques that are recommended to professionals as part of the "Step-Wise" interview. John Yuille used this procedure extensively in training professionals across the globe (see Yuille, Hunter, Joffe & Zaparniuk, 1993 for a description of Step-Wise). Similarly, in England and Wales, police and social workers have received training in a procedure recommended by the Home Office (The Memorandum of Good Practice, 1992; see section 17 of the target article) and this procedure forms the basis of the SI developed in the laboratory (Memon, Holley, Wark, Bull & Koehnken, 1996a; Memon, Wark, Bull & Koehnken, in press). Accepting the criterion of "current usage" as a marker of a valid comparison technique for the CI, then it is clear that the SI should not be discounted as a valid comparison.
5. The third point of importance to the adoption of the SI as a comparison group rests with the nature of the SI itself. There is a significant practical reason for comparing the CI with the SI which stems from the fact that the CI being practised in the field may actually be a watered-down version of the original interview as recommended by Fisher and Geiselman (1992), with only a few of the cognitive techniques being used. Work with police officers suggests that some practitioners are reluctant to use cognitive techniques due either to a lack of understanding, lack of resources for training, or time pressures (see Kebbell & Wagstaff, 1996; Turtle, 1995; Memon, Milne Holley, Koehnken & Bull, 1994). Readers should also note that there has been a misinterpretation of the points made by Kebbell and Wagstaff in their earlier commentary. Fisher (1996) cites Kebbell and Wagstaff in stating that "the CI is a technique that is currently used by practitioners in the target domain of interviewing". Kebbell and Wagstaff actually suggest that officers often don't have sufficient time to conduct a CI, and consequently, many don't actually use it (see Kebbell & Wagstaff, 1996, sections 3 and 4).
6. Finally, Fisher suggests that the SI fails to provide any theoretical insights into the workings of the CI. More than this, Fisher suggests that while the CI and SI differ procedurally by the inclusion of cognitive techniques in the CI, this addition of cognitive techniques affects non-retrieval factors (such as social dynamics) associated with the interview as well as retrieval factors (see Fisher, 1996, section 4). As such, Fisher argues that the SI is not a valid control group for isolating the factors that influence retrieval aspects alone.
7. This criticism can be dealt with easily, at one level. Quite simply, it cannot be disputed that the CI and the SI differ on the inclusion of cognitive techniques. Any differences in the effectiveness of the interviews must therefore be attributable to the cognitive techniques. This is a basic point of experimental design. As such, the SI is perfectly valid as a comparison for the CI for it permits an examination of the importance of the cognitive techniques that the CI-standard interview comparison cannot begin to address.
8. What may be open to debate, however, is the interpretation of the influence that these cognitive techniques have at a retrieval level and/or a non-retrieval level. This, however, is a psychological issue and should not be confused with the design issue above. Presumably, an appropriate experimental design could be devised to address Fisher's interpretational question. For instance, if we are interested in isolating the importance of cognitive techniques on retrieval alone, then we need to identify the components of the cognitive techniques that influence non-retrieval factors, and incorporate these components into a control interview. The comparison of the CI with this control would then enable a theoretically driven analysis of the role of cognitive techniques on information retrieval. Undoubtedly, a tighter comparison group than the SI could be devised. However, the benefit, not to mention the real world validity, of this tighter control seems marginal given that the CI-SI comparison already reveals little difference in interview accuracy. Furthermore, researchers would need to be much clearer than at present about exactly what the retrieval and non-retrieval factors actually are.
9. In summary, we disagree with Fisher's view of the SI as a poor comparison for the CI relative to the standard interview. The use of the SI in the field ensures that the CI-SI comparison is of practical value in assessing the CI relative to an existing interview technique. Furthermore, by holding social dynamics constant across the CI and SI, the SI permits the theoretically driven examination of the importance of the cognitive techniques that define the CI. In addition to these two points, the SI actually offers a stronger comparison than the standard interview by virtue of the greater degree of uniformity afforded by training. As such, we believe that the SI remains a useful comparison group against which the CI may be assessed. Together with the Fisher and Geiselman (1992) text, studies involving both the SI and the CI have advanced our understanding of the importance of the cognitive strategies, and of the interplay of cognitive and social factors in facilitating retrieval.
10. In addition to his concern about the standard of comparison when evaluating the CI, Fisher's commentary has raised another important issue -- that of the appropriate measure when assessing the accuracy of information elicited with the CI. Fisher echoes the concerns of Geiselman (1996) and highlights the need to attend to the error rate rather than the absolute number of errors during a given interview. Perhaps we have not been entirely clear in what we mean by accuracy. The term accuracy refers to the proportion of correctly or incorrectly recalled details. We agree entirely with Fisher (1996) that an increase in the absolute number of errors does not mean the CI is less accurate than the standard or structured interview. With the exception of one study, where the interview training was insufficient (Memon et al., 1994), we have not found the CI to be less accurate than a structured interview. Perhaps another way of saying this is that the information obtained with the CI is typically as reliable as that obtained with a comparable interview procedure.
11. However, our purpose in the target article was to caution interviewers that mistakes do occur when using any interview technique. In the field study reported by Fisher, Geiselman and Amador (1989) the researchers were fortunate enough to be able to corroborate almost every piece of information provided by witnesses through a comparison with the accounts of other witnesses. However, informal discussions with police officers suggest that it is often very difficult to assess the reliability of information in the real world. Accordingly, we cannot compute accuracy in real life interviews in the same way as we can in laboratory studies. Against this background of uncertainty for each piece of information elicited during the course of an interview, the presence of errors, even a relatively small number of errors, is of consequence (see Memon & Stevenage, 1996b). The reason that we drew attention to these errors was not to detract from the accuracy of the CI, but to remind practitioners that errors do occur with the CI. If we can encourage means of reducing the absolute number or errors elicited during the course of an interview, then we have the capacity to generate a technique that is even more efficient that the CI at eliciting both a full and accurate recall.
The authors gratefully acknowledge endless useful discussions with Dr Philip Higham.
Bull, R. & Cherryman, J. (1996). Helping to identify skill gaps in specialist investigative interviewing: Enhancement of professional skills. Final Report to the Home Office Police Research Group, London: HMSO.
Fisher, R.P. (1996). Misconceptions in Design and Analysis of Research With the Cognitive Interview. PSYCOLOQUY 7(35) witness-memory.12.fisher.
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.
Geiselman, R.E. (1996). On the Use and Efficacy of the Cognitive Interview. PSYCOLOQUY 7(11) witness-memory.2.geiselman.
Kebbell, M.R. & Wagstaff, G.F. (1996). Enhancing the Practicality of the Cognitive Interview in Forensic Situations: Commentary on Memon and Stevenage. PSYCOLOQUY 7(16) witness-memory.3.kebbell.
Memon, A., Holley, A., Wark, L., Bull, R. & Koehnken, G. (1996a). Reducing suggestibility in child witness interviews. Applied Cognitive Psychology, 10, 416-432.
Memon, A., Milne, R., Holley, A., Koehnken, G. & Bull, R. (1994). Towards understanding the effects of interviewer training in evaluating the cognitive interview. Applied Cognitive Psychology, 8, 641-59.
Memon, A. & Stevenage, S.V. (1996). Interviewing Witnesses: What Works and What Doesn't. PSYCOLOQUY 7(6) witness-memory.1.memon.
Memon, A. & Stevenage, S.V. (1996b). The Cognitive Interview in a Broader Context: Reply to Geiselman on Witness-Memory. PSYCOLOQUY 7(22) witness-memory.8.memon.
Memon, A., Wark, L., Holley, A., Bull, R. & Koehnken, G. (1996b). Interviewer behaviour in investigative interviews. Psychology, Crime and Law, 3, pp. 181-201.
Memon, A., Wark, L., Bull, R. & Koehnken, G. (in press). Isolating the effect of the cognitive interview techniques. British Journal of Psychology.
Shepherd, E. (1988). Developing interview skills. In, New Directions in Police Training, P. Southgate (Ed.), London: HMSO.
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.
Yuille, J.C., Hunter, R., Joffe, R. & Zaparniuk, J. (1993). Interviewing children in sex abuse cases. In G. Goodman & B. Bottoms (Eds.), Child victims, child witnesses: Understanding and improving children's testimony: Clinical, developmental and legal implications (pp. 95-115). New York: Guilford Press.