Cognitive Interview

Eyewitness information is the key element in solving many crimes, yet the police are often poorly trained in conducting information-gathering interviews, and they make avoidable mistakes. To rectify this situation, Ronald Fisher and Edward Geiselman developed the Cognitive Interview (CI) procedure to collect information from cooperative witnesses. The CI techniques are based on scientific principles of cognitive and social psychology and are intended to facilitate witness memory and communication between the witness and the interviewer. Laboratory and field tests have shown that the CI increases considerably the amount of information obtained from witnesses while maintaining high accuracy. This research paper describes the core elements of the CI, empirical tests to validate the procedure, and its various applications and limitations.

Police investigators depend heavily on eyewitness evidence to solve crimes, and they often bemoan the fact that witnesses do not provide as much information as the police expect. Some of this cannot be controlled, as when crimes occur quickly or under poor lighting conditions. Nevertheless, the police do have some control over witness recollection, specifically by the way they conduct interviews. Because many police investigators receive minimal training on how to conduct investigative interviews with cooperative witnesses, they conduct interviews intuitively and make avoidable errors. Studies of police interviews show that they

(a) ask too many closed-ended questions (e.g., How tall was the robber?) and too few open-ended questions (e.g., Describe the robber.), (b) often interrupt witnesses in the middle of their narrative descriptions, and (c) frequently ask leading questions.

To improve police interviewing procedures, Fisher and Geiselman developed an interview procedure that is based primarily on scientific, laboratory research in cognitive psychology (hence the name Cognitive Interview) and social psychology. The CI attempts to enhance witness recall by addressing three integral components of the interview: (a) developing effective social dynamics between the police interviewer and the witness,

(b) enhancing the witness’s memory retrieval and generally facilitating the witness’s and the interviewer’s thought processes, and (c) facilitating communication between the witness and the interviewer. The following is a thumbnail sketch of the CI’s core principles.

Social Dynamics

As in all small groups, the exchange of information depends on how psychologically comfortable the group members are with one another and each person’s expectations of his or her role in the group.

Developing Rapport

Witnesses, and especially victims, are often asked to give detailed descriptions of intimate, personal experiences to police officers, who are complete strangers. If anything, the police investigator’s formal appearance (badge, uniform, gun) may create a psychological barrier between the police officer and the witness. To overcome this barrier, police interviewers should invest time at the beginning of the interview to develop a meaningful, personal rapport with the witness, a feature often absent in police interviews.

Active Witness Participation

The witness has extensive first-hand information about the crime. Therefore, the witness, and not the interviewer, should be doing most of the mental work. In practice, however, police interviewers often dominate the social interaction with witnesses by asking many questions that elicit only brief answers. This relegates witnesses to a passive role, waiting for the police to ask questions. Interviewers can induce witnesses to take a more active role by (a) explicitly requesting them to do so, (b) asking open-ended questions, (c) not interrupting witnesses during their narrative responses, and (d) constructing the social dynamic so that witnesses perceive themselves to be the “experts” and therefore the dominant person in the conversation. The last point is especially important when interviewing children.

Memory and Cognition

Both the witness and the interviewer are engaged in demanding cognitive tasks: The witness is attempting to recall and describe in detail a complex event; the interviewer is listening to the witness’s response, generating and testing hypotheses about the crime, formulating questions, and notating the witness’s answers. Because these tasks are cognitively demanding, the witness’s and the interviewer’s cognitive resources must be used efficiently.

Context Reinstatement

Retrieving information from memory is most efficient when the context of the original event is re-created at the time of recall. Interviewers should therefore instruct witnesses to mentally recreate their cognitive and emotional states that existed at the time of the original event (What were your thoughts and emotions during the crime?).

Limited Mental Resources

Witnesses and interviewers have only limited mental resources to process information. Hence, their performance suffers when they engage in other difficult tasks concurrently. Interviewers can minimize overloading witnesses by asking fewer, but more open-ended, questions. This also makes the interviewer’s task easier by not having to formulate many questions. Interviewers can also promote a more efficient use of witnesses’ limited mental resources by minimizing physical (extraneous noises) and psychological distractions (direct eye contact) during the interview.

Witness-Compatible Questioning

Each witness’s mental record of an event is unique. Some witnesses may have focused on the perpetrator’s face, whereas others may have focused on the weapon. Interviewers should tailor their questions to each witness’s unique perceptions during the crime, instead of asking all witnesses the same set of questions. Interviewers often violate this rule by using a standardized checklist of questions for all witnesses.

The accessibility of event details varies during the course of the interview as the witness’s mental images change. Event details will be most accessible when they are perceptually related to the witness’s current mental image. Therefore, interviewers should be sensitive to the witness’s currently active mental image and ask questions that are compatible with that image.

Multiple and Varied Retrieval

The more often witnesses search through their memories of the crime, the more new details they will recall. Interviewers can make use of this principle by (a) asking witnesses to describe the critical event several times during the interview and (b) interviewing witnesses on two or more occasions. If witnesses attempt to recall the target event repeatedly, they should be directed to think about the event in various ways, since different retrieval cues will activate different aspects of a complex event. For instance, interviewers might encourage witnesses to describe the crime both visually (describe what the people and objects looked like) and temporally (describe the sequence of events).

Minimizing Constructive Recall

Witnesses may construct memories of a crime by incorporating information derived from other sources—for example, the media, other witnesses, or even the interviewer. Interviewers should therefore be careful about not leaking information to witnesses either nonverbally (e.g., by smiling or paying increased attention when the witness makes a particular statement) or verbally (by asking leading or suggestive questions, e.g., Was it a red car?).

Accuracy of Response

To promote high accuracy, interviewers should explicitly instruct witnesses not to guess; rather, witnesses should indicate that they “don’t know.” Interviewers should also refrain from applying social pressure on witnesses or otherwise encouraging them to answer questions whose answers they are unsure of. This is particularly important when interviewing children.


For police interviews to be effective, the investigators must communicate their investigative needs to the witness. Witnesses must also communicate their knowledge of the crime to the investigator. Ineffective communication will lead witnesses to withhold valuable information or provide irrelevant, imprecise, or incorrect answers.

Promoting Extensive, Detailed Responses

Police interviews require witnesses to describe people, objects, and actions in more detail than civilians normally do in casual conversation. To promote such extraordinary descriptions, police officers should convey explicitly their need for extensive detail, which they rarely do. Sometimes witnesses withhold information because they mistakenly believe that it is not relevant for a police investigation. To minimize witnesses’ withholding valuable information, interviewers should instruct witnesses to report everything they think about, whether it is trivial, whether it is out of chronological order, or even if it contradicts a statement made earlier.

Investigators often direct witnesses to provide relevant information by asking many specific, short-answer questions about investigatively relevant topics— for example, the perpetrator’s age, height, or weapon. This questioning style minimizes irrelevant information, but at the cost of minimizing unsolicited information and sometimes inducing incorrect responses. Rather than asking many specific questions, interviewers should explicitly instruct witnesses to generate descriptive narratives, without waiting for the interviewer to ask questions.

Code-Compatible Output

Interviewers and respondents often exchange ideas using only the verbal medium. Some people, however, are more expressive nonverbally, and some events are better described nonverbally. Ideally, the response format should be compatible with the witness’s mental record of the event. If an event is inherently spatial (e.g., the location of objects within a room), then witnesses should respond spatially—for example, by drawing a sketch of the room. In general, encouraging witnesses to sketch out the crime scene should promote more extensive recall.

Sequence of the Cognitive Interview

The CI follows a designated order intended to maximize the effectiveness of the individual techniques. The recommended sequence is common to many interviewing protocols in that it progresses from asking open-ended questions to more specific follow-up probes. The CI is divided into five sections: introduction, open-ended narration, probing for details, review, and closing the interview. The introduction establishes the appropriate psychological states and interpersonal dynamics to promote efficient memory and communication during the remainder of the interview. The open-ended narration allows the witness to provide an uninterrupted narrative of his or her recollection of the crime. The interviewer follows up by probing information-rich images, initially with framed, open-ended questions and then with more specific probes. When all the information has been collected, the interviewer reviews the witness’s statement to clarify any ambiguities and to resolve any contradictions. Finally, the investigator closes the interview by collecting official information (e.g., contact information) and encouraging the witness to contact him or her in the future.

Although this is the optimal sequence, interviews invariably deviate from this plan as unexpected conditions arise. In that regard, the CI is more of a general guideline for conducting an interview rather than a fixed recipe.

Empirical Testing to Validate the Cognitive Interview

The CI has been examined in approximtely100 laboratory tests, most of which were conducted in the United States, England, Germany, and Australia. In these tests, volunteer witnesses (typically, but not always, college students) observe either a live, nonthreatening event or a film of a simulated crime. Several hours or a few days later, the witnesses participate in a face-to-face interview, which is either the CI or a control interview. The control is either a “standard” police interview or a “structured interview,” which incorporates generally accepted principles of interviewing minus those techniques unique to the CI. The interviews are usually tape-recorded, transcribed, and then scored for the number of correct statements and incorrect statements. Across these studies, the CI has typically elicited between 25% and 100% more correct statements than standard or structured interviews. This effect is extremely reliable: Of the 55 experiments examined in a recent metaanalysis, 53 experiments found that the CI elicited more information than did the comparison interview (median increase = 34%). Equally important, accuracy was as high or slightly higher in the CI interviews (accuracy rate = .85) than in the comparison interviews (.82).

All the above studies were conducted in the laboratory, with nonthreatening events. Two other studies have examined the CI with victims and witnesses of real-world crimes. In both of these studies, one conducted in Miami and one in London, some experienced police detectives received training to use the CI and other experienced detectives did not receive such training. In both studies, the CI-trained police investigators elicited considerably more information than did the untrained investigators (approximately 50% more in the U.S. study).

Although most of the empirical testing has been conducted on normal, healthy adults, several studies have examined the CI’s effectiveness on unusual populations, including young children, the elderly, and those with cognitive deficits. Naturally, healthy college students remembered more than these other populations. However, the CI was equally effective with all the groups, enhancing their recollections by approximately the same amount. Some have questioned the advisability of using the CI with very young children, under the age of 5 years.

Most empirical studies have tested witness recall within a few hours or a few days of the critical event. Some studies, however, have shown the CI to enhance witness recall after several months, and one study even showed a very large benefit after 35 years.

The CI has been demonstrated to work effectively in a variety of investigative interviews in addition to criminal investigation—for example, accident or public health investigation. It has not, however, been effective in identification tasks: Witnesses given a CI prior to an identification test (e.g., lineup) were no more accurate than witnesses given a control interview.

Practical Issues in Cognitive Interviews

Given the success of the CI in laboratory and field experiments, how does it fare in real-world investigations? The CI has been used successfully to solve several cases, including a kidnapping, a politically motivated bombing, and child molestation. Recently, an investigator from the U.S. Bureau of Alcohol, Tobacco, and Firearms reported conducting a CI with a 38-year-old woman who had witnessed a homicide as a 5-year old child. The interview elicited scores of recollections, many of which were corroborated by police records established at the time of the crime (e.g., the location of objects and furniture at the crime scene).

Offsetting these successes, the British police reported that the complete CI was sometimes difficult to implement. They noted difficulty in communicating to witnesses some of the CI’s mnemonic instructions. Other police officers have reported that using the complete CI frequently requires more time than is available, and so they often use only some of the component techniques.

Other Investigative Tasks

Although the CI was developed initially to facilitate witness memory of a crime, the technique has been shown to be effective in other interview settings. Two such applications of the CI are interviewing suspects and debriefing jurors after deliberation.

Detecting Deception

Some research shows that the CI facilitates detecting whether a suspect’s testimony is truthful or deceptive. Two CI components that enhance detecting deception are asking open-ended information-gathering questions (vs. accusatory questions) and encouraging suspects to take an active role. These techniques generate longer responses from suspects, thereby permitting more opportunities to identify verbal and nonverbal cues to deception, and also allow interviewers to detect the different response strategies used by truth tellers and liars. In addition, asking suspects to describe events in different sequential orders (notably, reverse order) is particularly difficult for liars.

Debriefing Jurors

Reconstructing a jury deliberation session after a trial should assist attorneys to evaluate their trial strategies. A recent study examined the CI’s efficiency in reconstructing a related decision-making task (asking a small group of people to discuss business practices that entailed ethical decisions). The CI was modified slightly to account for group decision making (considering the social dynamics of the group). Compared with the conventional method of debriefing group members, the CI elicited considerably more information and at a very high accuracy rate. Interestingly, the CI also elicited extensive information about the individual members’ thought processes during the earlier decision-making task.

Componential Analysis

Although tests of the CI show that the technique, as a whole, is effective, only a few studies have isolated individual component techniques to determine which ones are effective. The results suggest that (a) each component contributes to the overall CI effect, but (b) the relative contribution of each component varies across conditions. For instance, context reinstatement is more effective when much time has passed between the original event and the interview, whereas nonverbal (code compatible) output is more effective when interviewing people with limited verbal skills.

Legal Challenges

Although the CI has been found reliably to enhance witness recollection, could it be unacceptable for forensic use? The following patterns of results suggest that the CI should be legally acceptable: (a) Cl-elicited recollections are as accurate as, or slightly more accurate than, recollections from conventional interviews; (b) the CI does not render witnesses overly suggestible to leading questions—if anything, witnesses are less suggestible when interviewed with the CI; (c) witness confidence and witness credibility are not affected by the CI; (d) CI interviewers are perceived to be less manipulative than conventional interviewers; and (e) there is no carry-over effect in a preliminary interview of the type of interview conducted (CI or conventional) on the witness’s later testimony.

There have been two court cases in which the CI was at issue. In a case heard by the National Court of Appeal in London (England), an earlier decision was overturned based on information collected from a witness who provided a very detailed account of the crime when interviewed with the CI. Although the

Court did not mention the CI in its ruling, the ultimate decision was compatible with the information elicited by the CI. The second case entailed a pretrial hearing in California, in which the prosecution used evidence that had been elicited by a police officer trained in conducting the CI. The defense attorney claimed that the CI was similar to hypnosis and that it promoted inaccurate eyewitness testimony. (As noted earlier, accuracy is equivalent or slightly higher with the CI compared with conventional interviews, the opposite of the pattern with hypnosis.) The judge ruled against the defense’s objection to the CI and permitted the CI-elicited testimony to stand.

Training in the Cognitive Interview

There is considerable variation across locations in the training the police receive to conduct interviews with cooperative witnesses. Several countries in Europe (England, Sweden, Norway) provide instruction in the CI as part of their basic training to all police investigators. Some regional police-training programs within the United States, Canada, and Australia also provide training in the CI, although (a) many police departments do not provide any training at all and (b) among those that do provide training in the CI, there is considerable variation in the quality. CI training is more standardized and more rigorous among some of the federal investigative agencies in the United States (e.g., FBI, National Transportation Safety Board [NTSB]). Adequate training in the CI requires, in addition to lectures and demonstration, ample opportunity for trainees to practice the techniques and receive critical feedback. Feedback from investigators has been very encouraging, especially with major, complex crimes and accidents, where the investigator has the luxury of time and resources to conduct thorough interviews.


  1. Fisher, R. P., & Geiselman, R. E. (1992). Memory-enhancing techniques in investigative interviewing: The cognitive interview. Springfield, IL: Charles C Thomas.
  2. Fisher, R. P., Geiselman, R. E., & Amador, M. (1989). Field test of the cognitive interview: Enhancing the recollection of actual victims and witnesses of crime. Journal of Applied Psychology, 74, 722-727.
  3. Fisher, R. P., & Schreiber, N. (2007). Interviewing protocols to improve eyewitness memory. In M. Toglia, R. Lindsay, D. Ross, & J. Reed (Eds.), The handbook of eyewitness psychology: Vol. 1. Memory for events (pp. 53-80). Mahwah, NJ: Lawrence Erlbaum.
  4. Geiselman, R. E., & Fisher, R. P. (1997). Ten years of cognitive interviewing. In D. G. Payne & R. G. Conrad (Eds.), A synthesis of basic and applied approaches to human memory (pp. 291-310). Hillsdale, NJ: Lawrence Erlbaum.
  5. Kebbell, M.-R., Milne, R., & Wagstaff, G.-F. (1999). The cognitive interview: A survey of its forensic effectiveness. Psychology, Crime and Law, 5(1-2), 101-115.
  6. Koehnken, G., Milne, R., Memon, A., & Bull, R. (1999). The cognitive interview: A meta-analysis. Psychology, Crime and Law, 5, 3-27.

Return to the overview of Eyewitness Memory in Forensic Psychology.