The best Forensic ArticlesEyewitness Memory: Can Suggestion be Minimized in the Investigative Interview?
By Wayne G. Whitehouse, PhD, DABPS, Emily Carota Orne, BA, and David F. Dinges, PhD
This article examines multiple sources of suggestive bias that may taint evidence obtained from forensic interviews with eyewitnesses and victims of crime. We present several research paradigms used to identify what constitutes undue suggestion, as well as the manner in which suggestive cues can be communicated and accepted by the witness as fact. Procedures that have been adopted as potential law enforcement tools to enhance eyewitness recall, such as forensic hypnosis and the cognitive interview, are submitted to special scrutiny and relevant cautions are raised. Finally, we provide specific recommendations to minimize suggestion in the investigative interview that are intended to expand upon current “good practice” guidelines.
One cornerstone of our justice system is the eyewitness report. Unfortunately, the substance and reliability of the report, conveyed in interviews and later in court, may be a direct function of the amount of undue suggestion that arises during the investigative interview process. Therefore, any scientific data that might establish a base for more careful investigative interviewing in order to avoid undue bias is of considerable relevance to law enforcement and the legal community.
Throughout this article we attempt to provide evidence of the pervasive nature of suggestion in everyday experience. Even in medical science, suggestion is taken into account when evaluating the effectiveness of new treatments. The classic “double-blind, placebo-controlled study” was developed to isolate the specific benefits of the treatment itself from those benefits that accrue from a hopeful patient’s confidence in his or her doctor as well as from the doctor’s belief in the efficacy of the prescribed treatment. A simple, elegant research strategy emerged: The treatment, say a medication in the form of a pink pill, would be administered to a random half of the patients enrolled in the study, while the remaining patients received an identically appearing pink pill that was actually inert (i.e., placebo). Not even the prescribing doctor knows whether the patient is receiving the actual medication or placebo. Therefore, the expectations of patients and their doctors for improvement of symptoms would be roughly equivalent regardless of whether the patients were receiving the actual treatment or the placebo. This essentially nullifies the influence of optimism, hope, and suggestion on the outcome, thereby permitting a clearer picture of the intrinsic efficacy of the medication. Unfortunately, the complications of suggestive bias on the products of human memory are not so neatly resolved.
The theoretical perspective advanced in this article developed from several decades of our laboratory’s research into a variety of investigative techniques. One of these, hypnosis, receives special attention because its hallmark feature is an increase in responsiveness to suggestion on the part of the hypnotized individual. Such situationally enhanced suggestibility, coupled with one’s native capacity to respond to hypnosis, allows hypnotized persons to modify their subjective experiences in accordance with the hypnotist’s instructions. This can lead to often remarkable positive outcomes, such as the suppression of pain during childbirth or dental procedures, but it is also responsible for instances of memory contamination and belief in false recollections, which can pose a serious threat to the very foundation of the justice system.
In order to identify what is unique about hypnosis, scientists long ago recognized the need to understand the role of suggestion in normal, everyday human interactions. Among the important lessons learned from this effort is that suggestion is ubiquitous in the social environment and that it doesn’t need to be intentional to have an impact. Even in the absence of explicit cues provided by the parties, people attempt intuitively to infer the motives behind another’s words or behavior. Martin Orne, who founded our laboratory in the 1960s, coined the term “demand characteristics” to describe the aggregate of elements in a social situation that provide information a person can use to draw inferences about someone else’s intentions.
Demand Characteristics: How do volunteers or witnesses bring their personal theories to bear on their reports? To illustrate the operation of demand characteristics, Orne and Whitehouse (2000) described an investigation in which research participants who volunteered for a “scientific study” were quite willing to engage in ostensibly meaningless activities for several hours—such as working rapidly and accurately to solve successive sheets of 224 addition problems, only to follow instructions to tear up each sheet into miniscule pieces before proceeding to the next!
When questioned later by a different researcher about their perceptions of the purpose of the study, participants invariably assigned great importance to their endeavors, viewing the experimental task as a test of endurance or something similar. Thus, the explicit communications consisted of instructions (a) to work quickly and accurately on a large number of arithmetic calculations (participant’s inference: “A test of my math ability”), and (b) to tear up your work and begin again (participant’s inference: “They don’t really care how many problems I got right—what a waste of time and effort!” and possibly, “But I’m supposed to continue, maybe this next one is the one that counts. Yeah, maybe they are trying to see what happens after I get discouraged.”). The participant continually develops, tests, and modifies hypotheses about the purpose of the study, which are intermittently or scarcely confirmed. When, after several hours repeating this pointless task, participants are queried about their impression of the study’s purpose, they still manage to rescue a meaningful hypothesis: “It was an endurance test.” On what evidence is this conclusion based? Among the demand characteristics operating here are subtle cues from the social context that enable critical assessments by the participant: “I volunteered to take part in a scientific investigation—it must be something important. These scientists are serious professionals. They know I have a busy schedule—they wouldn’t take up my time for frivolous purposes. They even offered to pay me at the end of the experiment.” In other words, the participants were responding to the implicit suggestion, communicated by a host of subtle demand characteristics, that their purpose was to make a contribution to advance scientific understanding and interpreted the importance of the research, as well as their roles, accordingly.
An altogether different scenario was depicted in a recent episode of a popular TV crime drama, but is still rife with suggestion created by demand characteristics. Here, one of the investigators, having just returned from collecting evidence at the crime scene, was consoling a female witness. After helping her to gain composure, he pulled out his digital camera and displayed the photograph of a male person-of-interest and asked essentially, “Was it him?” The witness affirmed that the person in the photo was the perpetrator. Notwithstanding whether the witness would have been able to identify the same person under different circumstances, the method used to obtain this identification was tainted by suggestion. Consider the demand characteristics that may have influenced the witness’s response: “I have just witnessed my friend’s murder. I am vulnerable and afraid for my own life. The police are the only ones who can protect me and I must help them. The investigator just came back from the crime scene and he has important evidence. He suspects a particular person was responsible; he has a photo in his camera’s memory; he hasn’t even printed it out yet; he needs me to identify this suspect.” The victim’s emotional state, her view of the police as authorities and the ultimate source of her protection, and the urgency with which the investigator sought identification of the suspect may have coalesced as a suggestive cue that led her to identify a particular individual. The level of suggestion could have been reduced had the investigator taken time to present the suspect’s photo in a less-biased photo identification array. Yet, even this might not sufficiently diminish the influence of suggestion, as witnesses may well infer that law enforcement agencies do not routinely maintain photo galleries of completely innocent citizens to use as foils in the identification process (Witness’s inference: “They are probably all guilty of something.”).
If this analysis leads the reader to the conclusion that suggestion is unavoidable in social interactions, then our point has been made. Nevertheless, it is unacceptable for investigators to be indifferent to this premise, because to do so is to diminish the standards of forensic evidence. Instead, the legal community needs to be aware of relevant research findings in the social sciences and be willing to revise best practice guidelines to remedy shortcomings in current investigative procedures. In the remainder of this article, we attempt to draw upon contemporary research that highlights problems and offers potential solutions to mitigate the effects of suggestive influence in forensic interviews.
Laboratory Research on the Vulnerability to Suggestion of Eyewitness Memory
A review of the literature reveals three distinct approaches to investigating the vulnerability to suggestion of eyewitness memory reports. One of these represents the field of applied memory research and relies upon the subtle introduction of leading or misleading post-event information by an investigative interviewer (Loftus, 1992). A typical experiment involves presenting a short film depicting some incident, such as a traffic accident. When subjects are later interviewed about what they had witnessed, the experimenter asks a question that includes a bit of explicit misinformation (e.g., substituting “yield sign” for “stop sign”) or a nuanced phraseology (e.g., “smashed into” for “bumped into”). While the theoretical basis for the effect remains controversial (e.g., McCloskey & Zaragoza, 1985), the evidence is clear that post-event misinformation conveyed, intentionally or inadvertently, by an interviewer stands a very good chance of being assimilated into the eyewitness’s account.
A second approach was developed in the context of basic memory research on the circumstances that lead to the production and acceptance of false memories (Roediger & McDermott, 1995). The standard paradigm involves the presentation of a number of conceptually related and nonrelated words, followed by testing for recognition of those same words when embedded in a list of nonpresented words, including one that is conceptually related, but nonetheless not previously presented (i.e., the critical lure). For example, a subject might see the words yawn, pillow, dream, pajamas, and bed interspersed among a list of nonrelated words like saxophone, computer, car, ocean, and leaf. The subsequent recognition test would include all of the previously presented words in addition to an equal number of nonpresented words, among which is the word sleep. The typical finding in this research is that subjects often mistakenly identify the conceptually related, nonpresented word (in this example, sleep) as having been presented in the earlier list—a case of false memory. Such research serves as a warning that eyewitness recollections can include schematically consistent details that were never perceptually experienced.
The third avenue of investigation into the impact of suggestion on eyewitness memory derives from the individual differences perspective and is best exemplified by the model of Gudjonsson and Clark (1986). These researchers endorse the objective assessment of witnesses’ “interrogative suggestibility” with respect to their “yield” and “shift” predispositions, corresponding, respectively, to their likelihood of succumbing to (mis)leading questioning (“yield”) or changing their testimony under pressure (“shift”). Their overall model takes into account the individual’s interrogative suggestibility, as measured by the Gudjonnson Suggestibility Scale (Gudjonnson, 1984)—which differs from hypnotizability (Register & Kihlstrom, 1988)—in addition to features of the interview context, such as the closed dyadic social interaction between interviewer and interviewee, procedural aspects of the interview, as well as the presentation, acceptance of, and behavioral response to, any “suggestive stimuli.” Much like contemporary hypnosis research, the Gudjonnson and Clark (1986) model considers the person-situation interaction as a fundamental determinant of the individual’s response. If the circumstances are stressful (as is typically the case), and the witness is uncertain about particular details, and is someone who is characteristically prone to “yield” or to “shift” his or her testimony, the situation is fertile for suggestion to contaminate the evidentiary record.
The scope of laboratory studies of eyewitness memory extends beyond the examples cited above. Drawing upon the same core paradigms—applied, basic, and individual differences approaches—researchers have been active in evaluating the pros and cons of various interview techniques adopted by law enforcement. Here we consider two such techniques, designed to enhance witness recall, that have emerged in forensic investigations over the past 40 years.
The Forensic Use of Hypnosis. Hypnosis gained a reputation as a useful agent of memory retrieval in the late nineteenth century with the publication of reports by Josef Breuer and Sigmund Freud, which cited its ability to access lost memories in psychoanalytic therapy—a claim that Freud later retracted. Nevertheless, it was not until the late 1960s (Harding v. State, 1968) that hypnotically elicited testimony was permitted as evidence in courts of law in the United States. This occasioned a proliferation of laboratory research aimed at identifying the mechanisms by which hypnosis might facilitate memory retrieval. Whereas some early studies reported positive or mixed findings, the more methodologically sophisticated research of the 1980s and 1990s revealed troublesome characteristics of hypnotically derived recollections (see Orne, Whitehouse, Dinges, & Orne, 1988). Among these were observations that (a) hypnosis influenced recall in an indiscriminate manner—posthypnotic testimony contained virtually equivalent amounts of new correct and incorrect information; (b) hypnosis increased the witness’s confidence that the newly accessed “memories” were accurate, thereby bolstering resistance to cross-examination; and (c) the expressed confidence of the previously hypnotized witness translated into greater perceived credibility of the testimony by jurors.
As this research unfolded, the judicial system responded with the imposition of various restrictions on hypnotically elicited testimony, ranging from full-blown “per se inadmissibility” (adopted by the majority of state Supreme Courts) to stipulating procedural safeguards, such as those recommended by Martin Orne in State v. Hurd (1981), to considerations of admissibility based on the “totality of circumstances” surrounding the hypnosis session as well as scientifically valid precedents (Daubert v. Merrell Dow Pharmaceuticals, Inc., 1993). The United States Supreme Court, in Rock v. Arkansas (1987), ruled that per se inadmissibility of hypnotically derived testimony could not be extended to the case of a previously hypnotized defendant in a criminal trial because it infringed upon the constitutionally guaranteed right to self-defense. Overall, scaling the legal hurdles established for the forensic use of hypnosis was becoming an impractical strategy for law enforcement, yielding to a search for viable alternative interview methods.
The Cognitive Interview. An apparently promising solution to the problem was embodied by the cognitive interview proposed by Geiselman, Fisher, Firstenberg, Hutton, Sullivan, Avetissian, and Prosk (1984), which consisted of a set of interview mnemonics derived from well-established basic research in human memory. The objectives are (a) to maximize the overlap of cues in the retrieval context with the memory trace of the target event, and (b) to encourage the witness to adopt multiple memory search strategies. Accordingly, witnesses are instructed to mentally reconstruct environmental circumstances associated with the incident, including their own thoughts, physiological, and emotional experiences; to report every detail that comes to mind, regardless of perceived importance; to recall events in a different order from the way in which they transpired; and to adopt the perspective of another witness in order to describe the events from that person’s point of view. A strong emphasis is placed on permitting the interviewee to engage in a comprehensive narrative with minimal interruption. Subsequently, and as appropriate, more specific prompts may be given to encourage the witness to elaborate on particular details. In recent years, the cognitive interview has been modified to include rapport-building strategies, visualization, and attention to the witness’s emotional circumstances and communication skills (Fisher & Geiselman, 1992; Geiselman & Fisher, 1997).
The cognitive interview has been widely adopted by law enforcement agencies and continues to be studied by researchers, throughout North America, Europe, and Australia. Its superiority to standard police interviews (in terms of eliciting a greater quantity of correct information) has been confirmed, but so has its tendency to moderately increase incorrect information. Its overall rate of accuracy (85%) was only slightly better than control interviews (82%) across the studies reviewed in a recent meta-analysis (Köhnken, Milne, Memon, & Bull, 1999).
For our purposes, the critical question is whether the cognitive interview can provide a law enforcement tool that escapes the problems associated with forensic hypnosis. The most straightforward way to address this issue is to compare the results of interviews involving hypnosis with those that employ the cognitive interview. Geiselman, Fisher, MacKinnon, & Holland (1985) reported such a comparison using films of simulated crimes as stimuli and experienced law enforcement personnel as interviewers. Here the original version of the cognitive interview was evaluated against a hypnosis interview and a standard police interview. Hypnosis and the cognitive interview produced comparable results, and both procedures yielded a greater amount of correct information than did the standard interview. Curiously, however, neither procedure resulted in more incorrect information than the standard interview. The latter finding is inconsistent with a large body of hypnosis research, which indicates that, typically, hypnosis produces increases in both correct and incorrect recall (for a review, see Orne et al., 1988).
In our attempt (Whitehouse et al., 2005) to replicate their findings, we recognized that Geiselman et al. (1985) used a measure of hypnotic ability that correlates very poorly with standardized instruments assessing hypnotizability (Orne et al., 1979). This makes it difficult to know how responsive their witnesses actually were to the hypnotic interview condition. Also, Geiselman and colleagues permitted their law enforcement interviewers to use hypnotic procedures with which they were familiar, rather than prescribing a standardized protocol. Our replication study also used a brief, filmed crime scenario, but all participants were screened for hypnotizability with a well-validated, tape-recorded assessment instrument, the Harvard Group Scale of Hypnotic Susceptibility, Form A (HGSHS:A; Shor & E. Orne, 1962). Standardized protocols were also adhered to in conducting our hypnosis, cognitive interview, and motivated repeated-recall control interviews.
Our findings differed from those of Geiselman and colleagues in several respects, the most important being that participants’ hypnotic ability, as measured by the HGSHS:A, was positively associated with the recall of erroneous and confabulatory (i.e., information that was never presented) details for subjects interviewed with hypnosis and the cognitive interview, but not for those interviewed in the motivated repeated-recall condition. This suggests that some components of the cognitive interview may tap hypnotic-like processes among hypnotizable individuals.
Let’s examine those features that hypnosis and the cognitive interview appear to have in common. Hypnosis relies on an ability to set aside current reality and to vividly imagine being in other times and places. Thus, hypnotized individuals can mitigate noxious events by engaging in imagery of pleasant experiences; they can strategically navigate the temporal realm by imagining themselves in the past or future. Hypnosis provides both license and a legitimate context for persons with the requisite cognitive ability to deploy fantasy to alter their in-the-moment circumstances.
Whitehouse et al. (2005) employed a hypnotic interview technique, popularized by Martin Reiser of the Los Angeles Police Department (Reiser, 1980), in which hypnotized witnesses are instructed to imagine they are watching the events of the crime unfold in a television documentary. They are told to use their imaginary remote control device to “zoom in,” “fast forward,” “reverse,” and “freeze frame” the TV footage in their “mind’s eye.” The metaphor provides a mechanism and rationale to review their recollections in different orders and amplify details in response to the investigator’s questions. Highly hypnotizable persons are quite able to do this; it is generally not as effective among comparatively less hypnotizable individuals. The cognitive interview uses similar instructions to guide memory retrieval, such as recalling events in different temporal sequences or recounting them from someone else’s point-of-view. Again, although less explicitly, witnesses are encouraged temporarily to discount their reality orientation and engage in fantasy. The ostensible difference between hypnosis and the cognitive interview is that typically the former employs a familiar, ritualized induction procedure to facilitate the transition from a normal waking state to the hypnotic condition. However, it is important to bear in mind that formal induction is not a necessary precondition for hypnosis to occur (Banyai & Hilgard, 1976; Estabrooks, 1948); hence this procedural distinction does little to undermine the proposition that the two interview procedures may involve similar cognitive processes.
An additional source of concern with the cognitive interview is the instruction to “report everything, regardless of its perceived importance to the investigation.” The stated intention of this instruction is to communicate to witnesses that they should not censor their reports with regard to confident recollections that they might feel are not relevant to the case. The danger, however, is that this may be misinterpreted to mean that they should include all information, even that about which they have little certainty (i.e., lower their subjective report threshold), which would engender the sort of problem that hypnosis creates—the potential inability of the witness to discriminate between factual and inaccurate recall. In response to this concern, Fisher and Geiselman (1992) now urge investigators to warn witnesses “not to guess or fabricate” (p. 41). Unfortunately, several survey studies of law enforcement personnel (Clifford & George, 1996; Kebbell, Milne, & Wagstaff, 1999) have determined that the cognitive interview is often modified in actual field use due, primarily, to time constraints, such that the prescribed version is not always the same as that which is ultimately adopted by forensic investigators.
In view of the difficulties associated with hypnosis and the cognitive interview, it is worth emphasizing that the motivated repeated-recall interview procedure used by Whitehouse et al. (2005) as a control condition in their study was found to be substantially less vulnerable to memory errors. It was based on extensive evidence that repeated efforts to recall previously presented information produced by corresponding incremental gains in memory output, referred to as hypermnesia (Erdelyi & Becker, 1974; Payne, 1987). No special intervention (e.g., hypnosis, guided imagery, cognitive interview) was necessary, only the willingness of the witness to really try several times to remember the same material. Our sole modification to this formula was to introduce intervening “distractor” tasks (see Whitehouse et al., 2005) designed to temporarily change the participant’s cognitive set, reduce fatigue, and to sustain repeated effort, which might not be as important (but should not be dismissed) in an actual forensic interview situation. The goal simply was to implement a control condition in which subjects were engaged in an equal number of recall attempts as participants in the hypnosis and cognitive interview conditions, with an equivalent motivation to expend recall effort.
Our experience with these interview procedures leads us to recommend (a) wholesale rejection of hypnosis due to its essential reliance on suggestion and fantasy; (b) further development and specific modification of the cognitive interview to minimize or replace the use of suggestive elements that can be activated by instructions to engage in visualization; and (c) emphasizing straightforward repeated entreaties to “try again” or “tell me everything you can remember,” as embodied by the motivated repeated-recall procedure. In a similar vein, Fisher, Brennan, and McCauley (2002) argue that additional refinements to the cognitive interview should now include explicit warnings to witnesses not to guess about details of which they are uncertain, and to avoid the “change perspective” technique when interviewing young children. Evidence from a recent study (Davis, McMahon, & Greenwood, 2005) suggests that certain mnemonic components of the cognitive interview (i.e., “change perspectives” and “change order”) can be replaced by additional free-recall opportunities— similar to the motivated repeated-recall procedure —without any significant reduction in its efficacy. Along with minimizing suggestive, fantasy-based aspects, these modifications to the cognitive interview by Davis et al. (2005) make it more feasible when time constraints are a concern.
Collectively, these developments contribute to a general procedural outline for investigative interviewing that can be used to minimize or, at the very least, identify instances of suggestive bias. The points described below are intended as a supplement to the comprehensive guide for law enforcement interviews published in 1999 by the National Institute of Justice (NIJ). The interested reader will, no doubt, recognize that elements of the cognitive interview, in its current state of development, are incorporated in the NIJ document. Rather than reiterate their recommended guidelines, our purpose is, instead, to highlight a few issues that sometimes create excursions from best practice standards, which, in turn, might open the door to suggestive bias.
Toward Minimizing Suggestion in Investigative Interviews
Of course, situational factors will always prevail and our recommendations here are intended to be generic. Implementation should be considered judiciously on the basis of such eyewitness factors as age, language skills, psychological disability, and relationship to any suspects in the case.
The scenarios described at the beginning of this article indicate that an important consideration should be the demand characteristics or contextual cues present in the interview setting. Law enforcement officials are perceived to embody vastly different personae, ranging from protective civil servants to fearsome bullies, with the power to incarcerate. Which role the interviewer adopts can have a considerable impact on the outcome of the investigation.
Most successful forensic interviews begin by developing rapport and ensuring the interviewee’s comfort. Part of this initial process involves conveying to witnesses that they, alone, possess definitive, critical information about the incident under investigation. The interviewer thereby establishes his or her role as a collaborator, rather than as an antagonistic authority figure. This cooperative relationship is then reinforced by encouraging witnesses to tell everything they can remember at their own pace and in as much detail as possible, while the interviewer refrains from interruption.
Prior to obtaining the witness’s narrative report of investigationally relevant aspects of the crime, however, it is often recommended (see National Institute of Justice, 1999) that the witness be encouraged to mentally reconstruct various environmental circumstances surrounding the event in question. These might include details such as the time of day, day of the week, weather conditions, prior activities of the witness, emotional state (e.g., upbeat, neutral, angry, or upset), physical state (e.g., activated, fatigued, or hungry), and how the witness happened upon the scene. The goal of this exercise is to mentally reinstate peripheral details that might serve as potential retrieval cues for other, more critically important, information to emerge during the witness’s free narrative. It is based upon a set of well-documented research findings, articulated by the encoding specificity principle (Tulving & Thompson, 1973), which imply that memory for an event is optimal when the conditions present at remembering match those present when the event was originally experienced. However, an important caveat that is often overlooked is that, in order to be maximally effective, the mentally reproduced contextual details must, in fact, include the critical retrieval cues pertinent to the incident under investigation. Thus, rather than running the risk that the effort to mentally reconstruct circumstances associated with the event of interest might lead to mismatched or irrelevant retrieval cues, in major forensic cases where the witness’s report may be critical, physically transporting the witness to the scene of the incident whenever possible is likely to prompt the most viable retrieval cues to aid the witness’s recollections of the event.
By far the most serious concern regarding the influence of suggestion on eyewitness memory arises when interview techniques that rely on fantasy and imagination are introduced in a situation in which the witness perceives salient demand characteristics to provide evidence consistent with a particular investigative theory. Recall from the examples at the beginning of this article that demand characteristics are often subtle and, unless anticipated and strategically managed by the investigator, they are apt to surface during the interview. This is a source of influence that appears to be ubiquitous in human social interactions. Ordinarily, honest, well-intentioned witnesses with relevant information to provide might stave off such pressure and, indeed, exert a countermanding influence (e.g., confidence) to accept their accounts as truthful. But the introduction of techniques such as hypnosis, visualization, and guided imagery can easily undermine the veridicality of testimony by rendering witnesses more compliant with demand cues from the investigator, while at the same time permitting them to abrogate personal responsibility for the accuracy of their statements. There is a well-documented scientific literature that demonstrates memory modification and confidence augmentation with hypnosis and suggestive imagery (e.g., Orne et al., 1988).
Recently it has been shown that instructions to close one’s eyes and merely imagine events that never happened increases confidence that such events really did transpire—a phenomenon known as imagination inflation (Garry, Manning, Loftus, & Sherman, 1996; Thomas, Bulevich, & Loftus, 2003). Research on this topic is noteworthy for several reasons. First, the imagination inflation effect has been observed in the general population and thus does not appear to be limited to witnesses who are extremely suggestible. Second, it is brought about by simple instruction, without any attempt to alter the witness’s state of consciousness as, for example, through a formal hypnotic induction procedure. Third, these findings have direct relevance to the development of the cognitive interview, where a similar strategy was recently incorporated in which witnesses are instructed to focus their attention on various sensory aspects of an incident by closing their eyes and visualizing the event (Fisher et al., 2002). Accordingly, the phenomenon of imagination inflation has serious implications for practices by law enforcement investigators and psychotherapists because it sheds light on the alarming ease with which events, once adamantly denied by the interviewee, might come to be confidently adopted as part of the individual’s autobiographical history in response to implicit or explicit suggestions to imagine they actually took place.
Thus it is clear that visualization and imagery based strategies to enhance recollection are fraught with the same problems that plague the forensic use of hypnosis, despite the absence of a specific induction ritual that is typical of hypnosis. This is why we strongly urge that specific modifications of the cognitive interview include the abandonment of such strategies as “visualization of sensory details,” “change perspectives,” and “recall events in a different temporal sequence.” If the witness did not actually experience events from someone else’s point-of-view or in reverse order, then to ask him or her to recount events from that perspective is literally asking the witness to engage in fantasy, i.e., to use imaginative skills to elaborate aspects of the event that could not otherwise be known. If it were the case that such an exercise had a benign impact on the witness’s memory and produced additional investigationally relevant details, then it certainly should be adopted. But the evidence strongly suggests that such techniques can sensitize witnesses to confidently report information about which they have little direct knowledge.
Finally, the need to authenticate the proceedings of investigative interviews cannot be overemphasized. As we have tried to demonstrate, suggestive bias emanates from a variety of sources—some ubiquitous, some unique, some inadvertent, and some even perhaps intentional. Regrettably, there is no known analog to the double-blind, placebo-controlled study—used by scientists to isolate the influence of suggestion from true treatment effects—available to investigative interviewers in their efforts to extract truthful evidence from eyewitnesses and victims of crimes. Instead, the reliability of evidence obtained from forensic interviews ultimately must be evaluated by jurors. Sworn testimony is often inadequate, particularly if there is any chance the witness has become convinced of the veracity of his or her account as a result of suggestive questioning by law enforcement officials. In such cases, the availability of some form of memorialized record of the interview session can be critical. At minimum the documentation could be in the form of a stenographic transcript provided by a disinterested third party, or, even more acceptable would be a well-labeled and well-archived audiotape recording. Ideally, in critical forensic cases, videotape that features the witness, a clock, and the investigator should be considered as it permits a full evaluation of the extent of duress, the demeanor of the investigator and the witness, as well as any verbal and nonverbal cues that might have been communicated with the potential to influence the witness’s report. Although every precaution needs to be taken to avoid the possible influence of suggestion on eyewitness memory in the first place, for justice to be served it is equally important that a reliable record of the means and methods used to acquire such evidence be made available to the courts.
References
Banyai, E., & Hilgard, E.R. (1976). A comparison of active-alert hypnotic induction with traditional relaxation induction. Journal of Abnormal Psychology, 85, 218-224.
Clifford, B.R., & George, R. (1996). A field evaluation of training in three methods of witness/victim investigative interviewing. Psychology, Crime and Law, 2, 231-248.
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S. Ct. 2786, 125 L. Ed. 2d 469. (1993).
Davis, M., McMahon, M., & Greenwood, K. (2005). The efficacy of mnemonic components of the Cognitive Interview: Towards a shortened variant for time-critical investigations. Applied Cognitive Psychology, 19, 75–93.
Erdelyi, M.H., & Becker, J. (1974). Hypermnesia for pictures: Incremental memory for pictures but not words in multiple recall trials. Cognitive Psychology, 6, 159-171.
Estabrooks, G.A. (1948). Hypnotism. New York: Dutton.
Fisher, R.P., Brennan, K.H., & McCauley, M.R. (2002). The cognitive interview method to enhance eyewitness recall. In M. L. Eisen, J. A. Quas, & G. S. Goodman (Eds.) Memory and suggestibility in the forensic interview (pp. 265-286). Mahwah, N.J.: Erlbaum.
Fisher, R.P., & Geiselman, R.E. (1992). Memory-enhancing techniques for investigative interviewing: The cognitive interview. Springfield, IL: Charles C Thomas.
Garry, M., Manning, C.G., Loftus, E.F., & Sherman, S.J. (1996). Imagination inflation: Imaging a childhood event inflates confidence that it occurred. Psychonomic Bulletin & Review, 3, 208-214.
Geiselman, R.E., & Fisher, R.P. (1997). Ten years of cognitive interviewing. In D.G. Payne & F.G. Conrad (Eds.), Intersections in basic and applied memory research (pp. 291-310). Mahwah, NJ: Lawrence Erlbaum Associates.
Geiselman, R.E., Fisher, R.P., Firstenberg, I., Hutton, L.A., Sullivan,, S., Avetissian, I., & Prosk. A. (1984). Enhancement of eyewitness memory: An empirical evaluation of the cognitive interview. Journal of Police Science and Administration, 12, 74-80.
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.
Gudjonnson, G.H. (1984). A new scale of interrogative suggestibility. Personality and Individual Differences, 5, 303-314.
Gudjonnson, G.H., & Clark, N.K. (1986). Suggestibility in police interrogation: A social psychological model. Social Behaviour, 1, 83-104.
Harding v. State, 5 Md. App. 230, 246 A.2d 302 (1968), cert. denied, 395 U.S. 949. (1969).
Kebbell, M.R., Milne, R., & Wagstaff, G.F. (1999). The cognitive interview: A survey of its forensic effectiveness. Psychology, Crime & Law, 5, 101-115.
Köhnken, G., Milne, R., Memon, A., & Bull, R. (1999). The cognitive interview: A meta-analysis, Psychology, Crime & Law, 5, 3-27.
Loftus, E.F. (1992). When a lie becomes memory’s truth. Current Directions in Psychological Science, 1, 121-123.
McCloskey, M., & Zaragoza, M. (1985). Misleading post-event information and memory for events: Arguments and evidence against the memory impairment hypothesis. Journal of Experimental Psychology: General, 114, 3-18.
National Institute of Justice. (1999). Eyewitness evidence: A guide for law enforcement. Retrieved July 15, 2007, from http://www.ncjrs.gov/pdffiles1/nij/178240.pdf.
Orne, M.T., Hilgard, E.R., Spiegel, H. Spiegel, D., Crawford, H.J., Evans, F.J., Orne, E.C., & Frischholz, E.J. (1979). The relation between the Hypnotic Induction Profile and the Stanford Hypnotic Susceptibility Scales, Forms A and C. International Journal of Clinical and Experimental Hypnosis, 27, 85-102.
Orne, M.T., & Whitehouse, W.G. (2000). Demand characteristics. In A.E. Kazdin (Ed.), Encyclopedia of psychology. Washington, DC: American Psychological Association and Oxford University Press.
Orne, M.T., Whitehouse, W.G., Dinges, D.F., & Orne, E.C. (1988). Reconstructing memory through hypnosis: Forensic and clinical implications. In H.M. Pettinati (Ed.), Hypnosis and memory (pp. 21-63). New York: Guilford Press.
Payne, D.G. (1987). Hypermnesia and reminiscence in recall: A historical and empirical review. Psychological Bulletin, 101, 5-28.
Register, P.A., & Kihlstrom, J.F. (1988). Hypnosis and interrogative suggestibility. Personality and Individual Differences, 9, 549-558.
Reiser, M. (1980). Handbook of investigative hypnosis. Los Angeles, CA: LEHI Publishing Co.
Rock v. Arkansas, 483 U.S. 44. (1987).
Roediger, H. L., III, & McDermott, K.B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 24, 803-814.
Shor, R.E., & Orne, E.C. (1962). The Harvard Group Scale of Hypnotic Susceptibility, Form A. Palo Alto, CA: Consulting Psychologists Press.
State v. Hurd, 432 A.2d 86 (N.J.). (1981).
Thomas, A.K., Bulevich, J.B., & Loftus, E.F. (2003). Exploring the role of repetition and sensory elaboration in the imagination inflation effect. Memory & Cognition, 31, 630-640.
Tulving, E., & Thomson, D.M. (1973). Encoding specificity and retrieval processes in episodic memory. Psychological Review, 80, 352-373.
Whitehouse, W.G., Orne, E.C., Dinges, D.F., Bates, B.L., Nadon, R., & Orne, M.T. (2005). The cognitive interview: Does it successfully avoid the dangers of forensic hypnosis? American Journal of Psychology, 118, 215-234.
Acknowledgment
Preparation of this article was supported in part by a grant from the Institute for Experimental Psychiatry Research Foundation.