Accuracy of Eyewitness Descriptions

Police investigators will frequently request that witnesses to a crime provide verbal descriptions of the alleged perpetrator. Such descriptions provide critical information that the police use throughout an investigation, from the identification of possible suspects in the vicinity of the crime, to the selection of photographs for mug books or lineup identification arrays, to the construction of sketches or composites that may be distributed to the general public. Although descriptions of persons are often accurate, they unfortunately also tend to lack sufficient detail to single out an individual suspect.

Quantity versus Quality of Person Descriptions

Numerous archival studies have examined the quantity and quality of person descriptions provided in real cases. On average, witnesses tend to provide between 7 and 10 descriptors, and these descriptors tend to be quite consistent (or congruent) with the defendant who is subsequently identified. Unfortunately, the vast majority of descriptors provided by witnesses are general, including characteristics such as gender, race, age, height, weight, build, and complexion. Aspects of the clothing worn by the perpetrator are also frequently mentioned, but such features provide only a brief opportunity for use in identifying a suspect in the immediate aftermath of a crime. More specific facial features (such as eye color, hair color or style, and face shape) are rarely mentioned by witnesses, and those that are included tend to focus on the upper portions of the face. Taken together, witnesses appear to provide an accurate general impression of the perpetrator but often fail to include more specific facial details. Laboratory studies of witness descriptions tend to concur with studies of real witnesses, indicating that although witnesses generally provide accurate descriptions, they rarely include descriptors that might be useful for individuating a target face.

Factors That Influence Description Accuracy

Research suggests that a variety of cognitive and social psychological factors can influence the accuracy of a witness’s description. First, encoding-based factors are those that occur around the time of the critical event when the witness interacts with or views the perpetrator. For example, low levels of illumination, greater distance between the witness and the perpetrator, a brief amount of time for viewing the perpetrator, the experience of stress or anxiety on the part of the witness (sometimes based on the presence of a weapon), and a witness under the influence of alcohol or drugs have all been shown to reduce the accuracy and completeness of person descriptions. Second, a subset of factors may occur between the time of encoding and retrieval of the description (i.e., during the retention interval) to influence the accuracy of a witness’s description. For example, longer delays between encoding and retrieval have been shown to significantly reduce the quality of descriptions provided by witnesses, and exposure to “misinformation” (as described later in this research paper) has been demonstrated to significantly impair a witness’s memory and thereby his or her person description. Finally, certain characteristics of the witness can influence the quality of his or her person description. In particular, adults tend to provide more detailed descriptions than do children, though few differences in the accuracy of person descriptions have been noted between these two populations. Similarly, young adults are superior at recalling person descriptions when compared with middle-aged and elderly adults. Interestingly, unlike the cross-race effect in face identification, few differences in accuracy have been noted when individuals attempt to describe faces of another, less familiar race or ethnicity.

Methods for Obtaining Person Descriptions

Interviewing techniques such as feature checklists, cued recall, and free-recall methods are well-established practices of investigators for eliciting person descriptions from eyewitnesses. Regardless of which technique is used, however, acquiring a complete yet accurate description has proven to be very difficult. Probably, the most common method for obtaining person descriptions is simply to ask the witness to freely describe what they remember about the perpetrator. While this free-recall technique regularly leads to highly accurate descriptions, critical details of distinguishing characteristics are often omitted from recall. Consequently, it is common practice for investigators to ask more direct, follow-up questions about specific features (e.g., “Do you remember if the man had facial hair?”) or to attempt to confirm the identity of a suspect that they have identified (e.g., “Did the man have short black hair and blue eyes?”). Studies suggest that such leading questions can be very dangerous in that they can “misinform” a witness’s original memory for the perpetrator and subsequently impair his or her ability to both provide an accurate description and identify the perpetrator. Research on feature checklist techniques similarly suggest that providing witnesses with numerous descriptors regarding a face can create confusion in memory and lead them to report the presence of features that they are actually unsure of. Finally, witnesses to a crime are often asked to describe the perpetrator many times over the course of an investigation. Research suggests that this process of repeated retrieval can have both positive and negative effects. On the positive side, repeatedly recalling information has been shown to lead to increases in recalled information and to offer some “protection” to the memory trace. Unfortunately, erroneous details generated during early retrieval episodes are also repeatedly recalled over time with increased confidence.

Of the attempts to develop an interviewing technique to maximize description completeness without sacrificing accuracy, the Cognitive Interview is perhaps the most well known. It has been shown to reliably improve the completeness of person descriptions in comparison with other “standard,” free-recall techniques. Unfortunately, some studies have suggested that the Cognitive Interview results in a slight cost in description accuracy in the form of increased errors. This has led some researchers to suggest that warning witnesses to be cautious in providing person descriptors may ultimately produce the greatest accuracy and simultaneously protect the witness’s memory from the confabulation of details.

The Description-Identification Relationship

It seems intuitive that an eyewitness who is capable of providing an accurate verbal description of a perpetrator would also be able to subsequently identify the perpetrator with greater accuracy; however, this seemingly obvious relationship between description and identification accuracy has not been demonstrated consistently in the research literature. For example, in what is known as the verbal overshadowing effect, researchers have demonstrated that asking participants to provide a verbal description of a face can actually impair their ability to subsequently identify that face from an array of similar photographs. In contrast, other studies have demonstrated that recognition of faces can be facilitated (or enhanced) by asking participants to provide a verbal description prior to test. A small body of literature has also assessed the specific relationship between verbal description and identification ability in memory for faces using a variety of measures of description quality, including indices of accuracy (the proportion of correct details reported), completeness (the total number of features reported), the frequency of correct and incorrect details that are reported, and the degree of congruence between the description provided and the face that is subsequently identified. Overall, there appears to be a small but reliable correlation between description accuracy and identification accuracy, and this effect appears to be particularly accounted for by the frequency of incorrect details that are generated in a description. Given the small size of the relationship between description and identification of faces, it appears possible that both memory tasks rely on a common underlying mental representation, yet also function on the basis of independent processing orientations (i.e., featural vs. holistic processing, respectively).

References:

  1. Meissner, C. A., & Brigham, J. C. (2001). A meta-analysis of the verbal overshadowing effect in face identification. Applied Cognitive Psychology, 15, 603-616.
  2. Meissner, C. A., Sporer, S. L., & Schooler, J. W. (2006). Person descriptions as eyewitness evidence. In R. Lindsay, D. Ross, J. Read, & M. Toglia (Eds.), Handbook of eyewitnesspsychology: Memory for people. Mahwah, NJ: Lawrence Erlbaum.
  3. Sporer, S. L. (1996). Describing others: Psychological issues. In S. L. Sporer, R. S. Malpass, & G. Koehnken (Eds.), Psychological issues in eyewitness identification (pp. 53-86). Hillsdale, NJ: Lawrence Erlbaum.

See also: