EFTA01657063.pdf

ResearchGate See discussions, stats.and author ;ogees let Ilics oubleation as: Mimi/Mem researchgale on/publicalem/20362017 Field Test of the Cognitive Interview: Enhancing the Recollection of Actual Victims and Witnesses of Crime Article w Journal of Applied Psychology • November 1989 co a toenail loan s 722 SI-gale tiro OTATIOXS READS 253 5,697 3 authors, including: Ronald Fisher Florida International University 130 PUBICA110XS 6.9119 CITATIONS rcop R. Edward Geiselman University of California, Los Angeles T9 PUS.10.11ONS 3.644 Crte. OWS SEE PROFILE SEE MERE Some of the authors of this publication are also working on these related projects: mow Metacogpitive Instruction and EncodingManipulations View project “neng lawns trio pep wnwbaclod ca 360•amintIDIA memo his toluene cenoxemeet oleic (lamas:led fik. EFTA01657063 Josmal al ADP'S Psr Ccpyright 1989 by the Aratrican Pgyebolnipail aortal:. Inc. PM. k 74. Mu 5. 722-727 0021.9010/10400 75 Field Test of the Cognitive Interview: Enhancing the Recollection of Actual Victims and Witnesses of Crime Ronald P. Fisher R. Edward Geiselman Florida International University University of California, Los Angeles Michael Amador Florida International University The Cognitive Interview was tested in the field to enhance the recollection of actual victims and witnesses of crime. The technique is based on laboratory-tested principles of memory retrieval, knowledge representation, and communication. Seven experienced detectives from the Metro-Dade Police Department were trained to use the technique and were compared with 9 untrained detec- tives. Before and after training, all detectives tape-recorded interviews with victims and witnesses of crime. The trained detectives elicited 47% more information after than before training, and 63% more information than did the untrained detectives. Overall collaboration rates (94%) were ex- tremely high and were equivalent for pre- and postuained interviews. Because the Cognitive Inter- view reliably enhances memory and is easily learned and administered, it should be useful for a variety ofinvestigathe interviews. Sanders (1986) asked sheriff's' deputies and detectives across several U.S. states have placed restrictions on the admissibility New Mork, "What is the central and most important feature of of hypnosis recall in a court of law. criminal investigations?" The majority of respondents an- In response to the need to improve police interview tech- swered, "Eyewitnesses." Nevertheless, few reported that they niques and to avoid the legal problems of hypnosis, Geiselman had any training in interviewing witnesses. Even though many and Fisher (Geiselman, Fisher, Cohen, Holland, & Suites, 1986; studies have sought to document and give theoretical explana- Geiselman, Fisher, Firstenberg, Hutton, Sullivan, Avetissian, & tions for the fallibility of witness memory (see Goodman & Prosk, 1984; Geiselman, Fisher, MacKinnon, & Holland, 1985) Hahn, 1987; Loftus, 1979; Yarmcy, 1979, for reviews), only re- developed a nonhypnotic interview procedure based on gener- cently has research been conducted on police interview tech- ally accepted scientific principles of memory.' The resulting niques to increase the completeness of a witness's report (e.g., procedure, called the Cognitive Interview, is a set of instruc- Wells, 1988). tions given by the interviewer to the witness at the beginning of One dramatic technique for eyewitness memory enhance- the interview. The goals of these instructions are (a) to encour- ment is hypnosis. Hypnosis has been reported to be useful in age the witness to reinstate the context of the original event and criminal cases, especially with traumatized witnesses (Reiser, (b) to search through memory by using a variety of retrieval 1980). Enhanced memory under hypnosis has also been found routes (see Geiselman et al., 1985, for specific details). The Cog- in some controlled laboratory experiments. In many studies, nitive Interview was compared with standard police interview however, researchers have found no memory enhancement with techniques in three laboratory experiments under highly realis- hypnosis. On the whole, the evidence about memory under hyp- tic conditions (e.&, using police films of simulated crimes). nosis is mixed (see Sanders & Simmons. 1983; Smith, 1983, for Overall, the Cognitive Interview elicited approximately 25%- reviews). Of greater practical consequence is that some re- 35% more information than did the standard police interview, searchers have concluded that hypnosis may distort the memory without generatingany more incorrect information (Geiselman process (see Geiselman & Machlovitz, 1987; Orne, Soskis, et al., 1984. 1985; Geiselman, Fishes Cohen, et al., 1986). We Dinges, & Orne, 1984). As a result of the inconsistency in the then refined the technique on the basis of insights gained from empirical literature, and as a general safeguard against the po- analyzing tape-recorded field interviews2 (Fisher, Geiselman, & tential problems encountered with memory under hypnosis, Raymond, 1987). In the revised version, which was evaluated in the present study. we approached the eyewitness's problem from the following three perspectives. representation of know!. This research was supported by a grant from the National Institute oflustice(USD1-854.1-CX-0053). This work was supported by an earlier grant from the National Insti- We would like to thank Brian Cutler for his helpful comments on an tute of Justice (USDJ-83-H-CX-0025). See Geiselman & Fisher 09861 earlier draft of this article. fora review of the research. Correspondence concerning this snide should be addressed to either 2 We sincerely appreciate the assigamx of Chief John S. Farrell, Bu- Ronald P. Faber, Department of Psychology. Florida International Uni- rcau Commander (Headquarters Detective Bureau). Lieutenant Ken versitg North Miami, Florida 33181, or R. Edward Geiselman, Depart- Russ. Sageant Jim Wander (Robbery Division), and the participating ment of INsystsulugy, University of California, Los Angeles, California detectives of the Metro-Dade Police Department. Dade County (Mi- 90024. ami), Florida. 722 EFTA01657064  COGNITIVE. INTERVIEW FIELD TEST 723 edge, memory retrieval, and communication. The following is couraged to continue trying to retrieve, even if they claim not a brief description of some of the core principles. A more com- to know a particular detail. plete and detailed description of cognitive interviewing is pro- vided in a short handbook by Fisher and Geiselman (in press). Witness-Compatible Questioning The primary issue of knowledge representation is that infor- mation about an event is represented at various levels of speci- Events are stored and organized uniquely for each witness. ficity (Fisher & Chandbx, 1988; Fisher & Cuervo, 1983). For Successful retrieval therefore reflects how compatible the ques- example, the representation of a bank robbery might be stored tioning is with the witness's unique mental representation. The at the very detailed level, including precise descriptions of the effective interviewer tries to tailor the interview to each witness event's actions and the robber's appearance and mannerisms. because a uniform style of questioning, asked of all witnesses and concurrently at the general level, that the event was "a bank alike, will not effectively tap idiosyncratic memories. Interview- robbery." Because the most valuable information, from the in- ers should be flexible and alter their approach to meet the needs cestigator's perspective. is stored at the detailed level, one of the of each witness rather than use a rigid, uniform style of ques- interviewer's goals is to maximize retrieval from the detailed tioning, thereby forcing witnesses to adjust their mental repre- level of representation and to minimize retrieval from the gen- sentations to the interviewer's questioning. eral level. Various cues (e.g., speech rate and word selection) can be used to recognize when retrieval is likely mediated by Specific Mnemonics the detailed level or the general level of description. In the ideal In addition to the general memory-retrieval principles men- interview, the interviewer guides the respondent to the detailed tioned, the Cognitive Interview includes a variety of mnemon- level of representation and then tries to maintain that level of ics to assist in retrieving specific pieces of information (e.g., description as long as possible. names, numbers. etc.). The primary ingredient in most of these The principal components of the Cognitive Interview arc mnemonics is to elicit partial information when the whole re- geared to enhancing memory retrieval by making witnesses sponse is unavailable. For example. if the witness cannot re- consciously aware of the events that transpired during the event. member a particular name, questions should be asked about The following four basic principles are used: event-interview specific, salient features of the name, such as ethnicity, length. similarity, focused retrieval, extensive retrieval, and witness- number of syllables, and so on. compatible questioning. The third component of the Cognitive Interview is geared to- ward facilitating communication of the witness's recollected Event-Interview Similarity events to the interviewer.' The communication principles are directed toward four goals, as follows: (a) assisting the witness Memory of an event, such as a crime, is enhanced when the to convert a conscious recollection into a detailed, elaborate psychological environment at the interview is similar to the en- response; (b) keeping the witness's statements "on target" that vironment at the original event (Reiser & Tulving, 1978). The is, relevant to the investigative needs of the interviewer; (c) facil- interviewer, therefore. should try to reinstate in the witness's itating the interviewer's comprehension and recording of the mind the external (e.g., weather). emotional (e.g., feelings of wimess's response: and (d) assisting the interviewer to under- fear), and cognitive (e.g., relevant thoughts) features that were stand the psychological needs of the witness. experienced at the time of the crime. Finally, a temporal sequence was developed which specifies the subgoals of the beginning, middle, and end of the interview Briefly, the interviewer's initial goal is to infer the respondent's Focused Retrieval mental representation of the event and then structure the re- Memory retrieval. like other mental acts. requires concen- mainder of the interview so as to be compatible with that repre- trated effort (Johnston, Greenberg, Fisher, & Martin, 1970). sentation. The interview is divided into five segments. The in- One of the interviewer's roles, then, is to encourage and assist troduction is used to establish rapport between the interviewer the witness to generate focused concentration. Any disruption and witness and to convey to the witness the appropriate psy- of the retrieval process, such as physical disturbances or inter- chological principles of memory. In the second stage, the inter- rupting the witness's narration, will impair performance. Fre- viewer encourages the witness to give an uninterrupted tuna- quently. witnesses will not attempt to search memory in a con- lion of the crime scene. This stage is intended more as a plan- centrated manner because of the additional mental "work" in- ning phase—for the interviewer to plan the strategy for the volved. In those instances, the effective interviewer must remainder of the interview—than as an information-collection encourage the witness to make the extra effort. phase. The middle of the interview is the information-gathering stage, when the interviewer guides the witness through various information-rich mental representations of the event. After Extensive Retrieval probing these mental representations, the interviewer reviews the witness's recollections. The interview is terminated for- in general, the more attempts the witness makes to retrieve a mally, but with a suggestion that prolongs its functional life. particular episode, the more information will be recalled (e.g., Reediger & Thorpe, 1978). Witnesses should therefore be en- couraged to conduct as many retrieval attempts as possible. I Although communication is not typically a problem in laboratory Many witnesses will terminate their retrieval attempts after the research, it can be a major hurdle in field interviews, in which victims first unsuccessful effort. In such cases, witnesses must be en- frequently arc extremely anxious and inarticulate. EFTA01657065 724 R. FISHER. E. GEISFLMAN. AND M. AMADOR The revised Cognitive Interview elicited approximately 45% interview The individual feedback session was an integral component more information than the original version, again, without elic- of the training, as many of the techniques explained in the lecture- iting any more incorrect information (Fisher, Geiselman, Ray- demonstration sessions were not fully implemented until after the feed- mond, Jurkevich, Warhallig„ 1987). Compared with similar back session. Because of the emergency nature of police work, changing schedules conditions in our earlier studies (Geiselrnan et al.. 1985: Geisel- and assignments, and mandatory court appearances. three members of man, Fisher, Cohen, et al., 1986), the revised Cognitive Inter- the trained group did not complete the entire training program. Our view elicited almost twice as much information as the standard results include only the seven detectives who completed the program. police interview. Having demonstrated reliably in the laboratory that the Cog- Posttraining Interviews nitive Interview can elicit more information than a standard police interview, we entered the last, and ultimately the most After the training phase, each of the seven trained and six untrained important, phase of the research that is, testing the Cognitive detectives tape-recorded 2-7 cases that met the aforementioned criteria. Interview in the field, with real victims and witnesses ofcrime. In all, 47 interviews were recorded, 24 by the trained group and 23 by As noted by Malpass and Devine (1980). the relevance of labo- the untrained group. As in the pretraining interviews, these interviews were primarily of victims ofeither commercial robbery or purse-match- ratory research will always be questioned unless it can be ap- ing. The posttraining interview phase took about 7 months to complete, plied to the real situation. Ultimately if the Cognitive Interview is to be applied outside the friendly confines of the laboratory it must be demonstrated to be effective in the real world. Our Analysis of-Interviews present research was geared toward that goal. All of the tape-recorded interviews were transcribed by a team of trained research assistants at the University of California. Los Angeles Method (UCLA). The transcribers were not told whether an interview was con- ducted by a trained or an untrained detective. The only identifying Interviewers marks on a cassette recording were the detective's name and case num- ber The transcriptions included only relevant, factual statements made Sixteen detectives from the Robbery Division of Metro-Dade Police by the eyewitness; none of the detective's questions were recorded. A Department. Dade County (Miami), Florida, were selected for the second group of research assistants, who were also blind to the condi- study. All of the detectives were experienced police officers, with a mini- tions. counted the number of relevant, objective statements made by mum of S years with the Robbery Division. the witness in each interview. Irrelevant statements (e4t, "I was going to work")arul opinionated statements (e.g., "The guY seemed nervous") Preliminary Interviews were not scored. The statements scored included primarily physical de- scriptions of the assailants and relevant actions: in addition. clothing, In the initial phase, all of the participating detectives were requested weapons, vehicles, objects taken, and conversations were reported. to tape-record their next several interviews, using standard interviewing procedures. The detectives overt asked to select the cases for recording Results using the following criteria: (a) Each case was to be serious enough so that ample time and resources were availabie if necessary, to conduct a The effectiveness of the Cognitive Interview can be examined thorough interview; (b) at least one victim or witness had a decent in the following two ways: (a) by comparing the number of facts chance to observe the suspect a suspects and the event; and (c) each elicited before and after training for the detectives who com- interviewed victim or witness had to be reasonably fluent in English pleted the training program and (b) by comparing the number and coopttative. Cases to be eliminated included those in which the interview was conducted more than a few days after the crime, when of facts elicited by the trained versus untrained detectives. As the witness was intoxicated, when the suspect was clearly known to the Table I shows, the Cognitive Interview was found to be effective witness• or when a suspect had been detained for identification! in both the before-after comparison and the trained-untrained The preliminary phase of interviewing took 4 months to complete, groups comparison. As a group, the seven trained detectives with each detective recording 5-7 him:views In all. 88 interviews were elicited 47% more information after than before training, F(1, recorded, primarily with victims of commercial robbery or purse- 6) = 12.66, MS, - 45.49, p < .05. Of these seven detectives, snatching. On the basis of the amount of information gathered in these six elicited more information after than before training (34%- preliminary interviews and the recommendations of the detectives 115% improvement). Only one detective did not do appreciably commanding officer, two equivalent groups of detectives were formed. better after than before (23% decrement). Not coincidentally, One group was trained on the Cognitive Interview; the other group was an analysis of the posttraining interviews showed that he was untrained and served as the control. the only one of the seven detectives who did not incorporate the recommended procedures into his posttraining interviews. Training in the Cognitive Interview The comparison between the trained and untrained detec- The training vast conducted in four 60-min group sessions, including tives is shown in the Training x Phase interaction, Fj1, It = lectures describing various components of the procedure and demon- 9.01, MS, = 27.04, p < .05. Planned comparison tests indicated strations of good and poor interviewing techniques. The schedule of that the trained and untrained groups were equivalent before topics was training. F(1. II) < 1, MS, rs 88.16, but that the trained group Session I: Overview and principles of cognition Session 2: Specific interviewing techniques to enhance memory Session 3: Enhancing eyewitness-interviewer communication ' When police have detained a suspaa for identification, interviews Session 4: Temporal sequence of the Cognitive Interview. with eyewitnesses who can probably make an identification are some- After the fourth session, each detective tape-recorded a practice inter- times less detailed than they would be otherwise, as the police are con- view in the field and received individual feedback on the quality of his caned primarily with securing a positive identification. EFTA01657066  COGNITIVE INTERVIEW FIELD TEST 725 elicited considerably more (63%) information after training. Table 2 F(I.11) = 4.84, W e n 157.46. p < .05. Comparison ofPre- and Posurained Detectives' Interviews Because these analyses were conducted on only a limited with Uniformed Officers'Reports number of cases (24 posttrained interviews), the possibility ex- Relation between Detectives' and Uniformed ists that these few cases were unrepresentative of the entire sam- Officers' Interviews ple of cases. Perhaps the 24 posttrained cases involved crimes that occurred under better observing conditions, or perhaps Research phase Same Different New 'Total these particular witnesses had unusually good verbal skills. Al- though this seems unlikely because no special instructions were Before training 12.76 1.45 2127 35.48 After training 13.68 1.68 34.45 49.82 given to the detectives when tape-recording posttraining inter- views, we examined the possibility that these were particularly Note. Cell entries are the number ofelicited facts. easy interviews to conduct. In each of the eases analyzed, the eyewitness was interviewed by a uniformed police officer before being interviewed by the detective. Presumably, "easy" inter- only the amount of information, but also its accuracy. To what views, cases involving witnesses with good verbal skills or good degree might the additional information elicited by the Cogni- viewing conditions, should be apparent from the amount of in- tive Interview simply reflect a lower response threshold and a formation in the uniformed officer's initial interview Thus, in concomitant decrease in accuracy? In previous laboratory stud- "easy" interviews, witnesses should generate more information ies, we found no differences in the accuracy rates of the Cogni- for both the follow-up detective and the uniformed officer; tive Interview and standard police interviews (Geiselman et at. whereas, in "difficult" cases, witnesses should generate less in- 1985; Geiselman, Fisher, MacKinnon, et al., 1986). Approxi- formation in both the detective's and the uniformed officer's mately 85% of all the statements elicited were correct, in all interviews. As an unbiased measure of the quality of the detec- conditions. In a field study, there are obviously no data about tive's interview, we scored the transcripts in terms ofhow much accuracy because one cannot determine exactly what tran- additional information the detective elicited compared with the spired during the crime. We therefore estimated accuracy by uniformed officer. Each statement in the detective's interview comparing a witness's report with another reliable source of was categorized as being either the same as found in the uni- information. In 22 cases there was another witness or victim, in formed officer's report (some), containing new information not one case a confession, and in one case a film from a hidden described in the uniformed officer's report (new), or being camera. When the source was another witness or victim, the different from that reported by the uniformed officer (different). interview was almost always conducted by someone other than In all, we examined 29 interviews conducted by detectives be- the detectives participating in the study (typically, a uniformed fore training in the Cognitive Interview and 22 conducted by police officer) and within a few minutes after the crime. Obvi- detectives after training.' The results, which are shown in Table ously, corroboration with another witness measures something 2, mirror the analysis of total number of facts. Overall, more other than accuracy, as witnesses can corroborate one another's information was collected in the posttrained interviews (49.82 report, yet both he inaccurate. Nevertheless, we can still expect facts) than in the pretrained interviews (35.48), RI, 49) = 3.55. corroboration to be correlated with accuracy, and in the field, NSr — 81.85, although the effect here was only marginal. .05 < with no perfect replica of the crime, corroboration is our best p <.10. The effect of training interacted with type of fact (same, estimate of accuracy. In the 24 cases with corroborating evi- new, different), F(2, 98) - 3.96, M.5., - 55.57, p < .05. The dence (16 by pretrained detectives and 8 by posttrained detec- difference between pre- and posttrained interviews was ob- tives), there were a total of 325 corroborable statements. Over- served only for new information collected, facts that the uni- all, 94% were corroborated.' More important, the corrobora- formed officer had not uncovered. There were no differences tion rates were equivalent for the pretrained (93.0%) and for the same and different information. Theoretically; the detec- posttrained (94.5%) interviews. The similarity of the corrobo- tive's role is to elicit additional information from that collected ration rates for the Cognitive Interview and the standard police by the uniformed officer. Information that duplicates the uni- interview duplicates the laboratory findings with accuracy rates formed officer's report (same) provides no new insights for the and again suggests that the added information elicited by the police investigation, and different information just casts doubt Cognitive Interview does not come at the expense of increasing on the reliability of the witness or investigation procedures. incorrect information. That the superiority of the posttrained group occurred only for new information testifies to its practical utility. 5 Not all of the detective interviews could be compared with the uni- As with the laboratory• studies, we were concerned with not formed officers' reports, as some reports were inaccessible. 6 Note that the corroboration rate is extremely high in comparison with the accuracy rates reported in typical laboratory studies. Similarly, Table 1 high accuracy rates were reported in field studies by Wilk and Cutshall Number ofFacts Elicited by Trained and UntrainedDetectives (1986) and Wilk and Kim (1987). Although they are not definitiw, it is interesting that the accuracy-corroboration rates in the three field Training group studies ofeyewitness memory were considerably higher than their labo- ratory countapartt If this diffaencc between laboratory and field stud- Research phase Trained Untrained ies continues to appear, one may question the validity of describing in court the accuracy rates found in the laboratory as evidence of the gen- Before training 26.83 23.75 eral unreliability of eyewitness testimony in field cases (McCloskey & After training 3937 24.21 Egeth, 1983). EFTA01657067 726 R. FISHER, E. GE1SELMAN, AND M. AMADOR Because the Cognitive Interview is more complex and en- tive Interview compare with hypnosis? Although we have never courages more extensive and focused retrieval. we expected it directly compared the current version of the Cognitive Inter- to take longer to conduct than the standard police interview view with hypnosis, Fisher, Geisclman, Raymond, Jurlcevich, Such differences have been found in previous laboratory studies and Warhaftig (1987) found that in similar observing condi- (Fisher, Geiselman, Raymond, Jurkevich, & Warhaftig, 1987; tions across two experiments, the Cognitive Interview elicited Geiselman et al., 1985); however, in neither of these two studies 33.4% more correct information than did hypnosis. The Cogni- was the extra time responsible for the superiority of the Cogni- tive Interview also is not beset with the problems sometimes tive Interview. In the present study, the interview times were found with hypnosis. The Cognitive Interview does not lead to surprisingly similar for the various interviews. The posttrained increased error rates and does not render respondents hypersug- interviews of the trained detectives (11.47 min) were not reli- gestible to leading questions; if anything, it mitigates the effect ably longer than their pretrained interviews( l0.65 min) or than (Geiselman, Fisher, MacKinnon, et al., 1986). Yuillc and Kim those conducted by the untrained detectives (9.05 min). both (1987) recently analyzed several police hypnosis interviews and h's< I, = (1, 6) and (I, II), and Atces = 10.587 and 21.295, found that interviewers frequently used some components of respectively. Of the sewn trained detectives, four took more the Cognitive Interview. Yuille and Kim concluded that "the time to conduct posttrained interviews than pretrained inter- advantage of hypnosis found in the present study has nothing views, and three took less time. In this study, therefore, the ob- to do with hypnosis. . . . [Instead) the cognitive interview is the served superiority of the Cognitive Interview over the standard 'active' memory component ofhypnosis" (p. 427). Because the police interview was not due to differences in interview time. Cognitive Interview has been a reliable technique to enhance recall, yet has none of the negative consequences of hypnosis, Discussion it has been suggested as a preferable alternative (Delfenbacher 1988; Orne, as cited in Commonwealth v. Di Nicole', 1985), and The overall pattern of data—an increase in the amount of one that should be more acceptable by the courts (Yoffie & investigatively relevant information accompanied by extremely Kim, 1987). high corroboration ratcs--provides strong support for the In addition to its technical merit, the Cognitive Interview has effectiveness of the Cognitive Interview in field investigations. considerable practical utility; that is, (a) the technique can be Training in the Cognitive Interview enhanced performance in learned within a few hours, (b)it requires little theoretical back- both the before-after and the between-groups comparisons. Six ground or previous training, and (c) it is easily administered. of the seven detectives improved with training; only the one de- Finally, because the technique is based primarily on proper tim- tective who did not use the technique as intended did not im- ing, sequencing, and phrasing of questions, the respondent prove. In the between-groups comparison, the two groups of should not perceive that any special intervention is occurring. detectives were equivalent before training, as judged by their As such, eyewitnesses should have no apprehensions about par- supervisors and by empirical observations. After training, the ticipating in this form of interview. trained detectives elicited 63% more information. That the Although the Cognitive Interview is an effective investigative effectiveness of the training was comparable for the before-after instrument. there are a few limitations. First, its usefulness may and between-groups comparisons suggests that the conclusions vary from one event to another Its primary contribution for are unlikely to be a product ofanything unique to either design. police will be in cases such as commercial robbery or battery, Furthermore, the magnitude of the effect (50%-60% improve- in which the bulk of the evidence comes from eyewitness re- ment) is reasonably dose to what we would expect on the basis ports, as opposed to crimes that rely more heavily on physical of the previous laboratory findings, in which the revised tech- evidence. Second, the Cognitive Interview can be used only with nique was 45% more effective than the original (Fisher Geisel- cooperative witnesses. Witnesses who wish to withhold infor- man. Raymond, Jurkevich, & Warhaftig, 1987), which, in turn, mation intentionally will not be "broken" by the Cognitive In- was 25%-35% more effective than standard police interviews terview. Third, although our study found no differences be- (Geiselman et al., 1985; Geiselman, Fisher, Cohen, et al., 1986). tween the time taken to conduct the Cognitive Interview and a The similarity of the corroboration rates associated with the standard interview; we expect that the Cognitive Interview takes Cognitive Interview and with standard police interviews also somewhat longer. It can be used to greatest effect, therefore, mirrors the findings from our laboratory studies, in which aeon- when there is ample time to conduct the interview. Finally. the racy rates were equivalent. In sum, the Cognitive Interview in- Cognitive Interview requires considerable mental concentra- creased the effectiveness of investigative interviews without any tion on the part of the interviewer He or she must make more apparent negative consequences. on-line decisions and show greater flexibility than is typically How does the Cognitive Interview compare with other mem- demonstrated in standard police interviews. In that sense, it is ory-enhancing techniques used by investigative interviewers? probably more difficult to conduct the Cognitive Interview than Historically, little training has been provided to law-enforce- the standard interview. As with other skills, however, with prac- ment interviewers to enhance the recollection of cooperative tice, many of the resource-demanding mental operations re- witnesses. Similarly, there is no formal training in memory-en- quired initially will he handled automatically. hancement techniques for law students or attorneys to conduct Because this version of the Cognitive Interview contains sev- effective interviews with clients, even though the fact-finding eral components, we cannot be sure which of the suggested stage is often critical to successful legal inquiry (Fisher, 1986). techniques are primarily responsible for its overall effectiveness. We know of no formal technique other than hypnosis that ap- Some components may have only marginal value. Others may pears regularly in the training of investigative interviewers, at vary depending on the specific features of the interview, for ex- least for memory-enhancement purposes. How does the Cogni- ample, whether the witness is a child or an adult and whether EFTA01657068  COGNITIVE INTERVIEW HELD TEST 727 interviews are conducted immediately after the crime or after a borne outbreaks of disease. Unpublished manuscript, Florida Inter- long delay. Sane components of the Cognitive Interview have national University, Miami. been isolated and demonstrated to be effective in an eyewitness Flea A.. & Ildving, E. (1978). Retrieval independence in recognition task (e.g., witness-compatible questioning Fisher & Price- and rocall. PsychologicalReview 85. 153-171. Geiselman, R. E.. & Fisher, R. P. (1986). Interviewing victims and wit- Roush, 1987; context reinstatement Geist!man, Fisher, Mac- nesses of crime. U.S. Depannient ofJustice. Research inMid 1-9. Kinnon, et al., 1986). we invite other researchers to tease apart Geiselman, R. F.., Fisher, R. P., Cohen, G., Holland, H., & Surtes, L. the various components (more explicitly described in Fisher & (1986). Eyewitness responses to leading and misleading questions un- Geiselman, in press) to determine their relative efficacy and to der the Cognitive Interview. Journal ofPolice Science and Adminis- help refine the technique even further. At this time, given (a) its tration. le. 31-39. proven record in the laboratory and in the field. (b) the ease of Geiselman, R. E., Fisher, R. P. Firstenbag, I., Hutton, L. A., Sullivan, learning and implementation, and (c) the positive feedback we S. J., Avetissius I. V., & Prosk, A. L. (1984). Enhancement ofmewit- have received from detectives who have used it, we recommend ness memory: An empirical evaluation of the Cognitive Interview. that the Cognitive Interview be incorporated into the standard JournalofPoliceScierwe ana'Administration 12, 130-138. Geiselman, R. E., Fisher, R. P., MacKinnon, D. P., & Holland, H. L training program ofall investigative interviewers. (1985). Elessimess memory enhancement in the police interview: As a final note, although we have examined the Cognitive Cognitive retrieval mnemonics versus hypnosis. Journal of Applied interview as a method to improve criminal investigations from Psychology 70.401-412. the police's (or prosecutor's) perspective, the technique can be Geiselman, R. E. Fisher, R. P., MacKinnon, D. P.. & Holland, H. L. used equally well to help defend innocent suspects. The Cogni- (1986). Enhancement of eyewitness memory with the Cognitive In- tive Interview simply facilitates the recollection of relevant evi- terview. Amen= Journal ofPsychology 99, 385-401. dence; whether the gathered evidence serves to acquit or convict Centime's R. E, & Machlovitt H. (1987). Hypnosis memory recall: is immaterial. In a similar vein, the technique should be useful Implications for forensic use. American Journal ofForensic Psychol- for civil as well as criminal investigations. In its broadest scope, ogy, .5, 37-47. because the Cognitive Interview is based on principles ofeosin- Goodman, G. S., & Hahn, A. (1987). Evaluating eyewitness testimony. In 1. B. Werner & A. K. Hess (Eds.). Handbook offorensicpsWholoS31 tion and is not limited by any specific content area, it should be New NOE: Wiley. useful for a wide range of investigations, for example, develop- Johnston, W. A., Greenberg, S. N., Fisher, R. P.. & Martin, D. W. ing medical histories, clinical interviews, journalistic inter- (1970). Divided attention: a vehicle for monitoring memory pro- views, accident investigations, oral histories, and so forth. As cesses. Journal ofExpenmental Psychology 83. 164-17 I . evidence of this generalizstbility, Fisher and Quigley (1988)have Loftus, E. F. (1979). Eyewitness testimony Cambridge, MA: Harvard shown that a variant of the current technique can be used in University Press. epidemiological investigations to trace foodbome outbreaks of Malpass, R. S.. & Devine, P G. (1980). Realism and eyewitness identi- disease. fication research. Law andHuman Behavior: 4. 347-358. McCloskey, M.. & Egeth, H. (1983). What can a psychologist tell a jury/ American Psychologist 38. 550-563. References Orne, M. T., Soskis, D. A., Dinges. D. F., &Gine, E. C (1984). Hypnoti- cally induced testimony. In G. L Welk & E. F. Loftus (Eds.), Eyewit- Commonwealth v. Di Nicola. Pa. (1985). ness testimony.' Psychological perspectives. New York: Cambridge Deffenbasher, K. A. (1988). Eyewitness research: The next ten years. University Press. In M. Gruneberg, P. Morris, & R. Sykes (Eds.). Pretrial aspects of Reiser, M. (1980). Iforilbook ofinvestigativehypnosis. Los Argyles CA: memory: Current research andissues. New York Wiley. LEHI. Fisher, R. P. (1986). Client memory enhancement with the cognitive Roediger, H. L.. & Thorpe, L A. (1978). The role ofremnantin pro- interview. Florida Bar Journal. 60(11). 53-56. ducing hypermnesia Memory & Cognition, 6.296-305. Fisher. R. P., & Chandler. C. C. (1988). Independence ofthente-cuedarrl Sanders, G. S. (1986). The usefulness of eyewitness researchRent the episodically cuedrecall. Unpublished manuscript perspective indict investigators. Unpublished manuscript, State Fisher. R. P., & Cuervo. A. (1983). Memory for physical features of University of New York at Albany. discourse as a function of their relevance. Journal ofExperimental Sanders. G. S.. & Simmons. W. L. (1983). The use of hypnosis to en- Psychology: Learning. Ffentory andCognition. 9. 130-138. hanceeyewitnessaccuracy: Does it work?JoirmalofAppliedPsychol- Fit. R. P., & Geiselman, R. E. (in press). Memory-enhancemeru tech- ogy 68.70-77. niquesfar investigative interviewing: The Cognitiveinterview Spring- Smith. M. (1983). Hypnotic memory enhancement of witnesses: Does field, IL: Charles C. Thorns,- it work? PsychologicalBulletin, 94, 387-407. Fisher. R. F. Geiselman, R. E.. & Raymond, D. S. O987). Critical anal- Welb, G. (1988). Eyewitness identification. A system handbook. To- ysis ofpolice interview techniques. Journa