Snap Shot
- The question: Can describing a culprit’s face in words make a witness less able to recognize that face later in a lineup?
- What the research shows: Under some conditions, yes. A meta-analysis and a 31-laboratory replication found a reliable but modest drop in correct identifications after a description, larger when the description was given after a delay following the event and the identification test followed soon after the description. Recent work suggests that describing can both reduce memory-based discrimination and make witnesses more reluctant to choose.
- What it does not show: The research does not show that taking descriptions is harmful practice, that a described-then-identified face was misremembered, or that any particular identification is wrong. On average the effect is modest, varies with timing and task, and in some conditions disappears or reverses.
- The law: The accuracy of the witness’s prior description is one of the Neil v. Biggers reliability factors, applied in federal court, in courts-martial under Mil. R. Evid. 321, and in Arizona at a Dessureault hearing. Expert testimony on the effect is governed by Rule 702 (federal, military, and Arizona), Houser, and Chapple.
- Issues for each side: Prosecutors should document when and how a description was taken and what came next, without dismissing the effect; defense counsel can raise the sequence and timing of description and identification, without presenting a small, condition-dependent effect as proof of error.
Investigators ask witnesses to describe the person they saw because a description is needed to search for a suspect, to broadcast a lookout, and to preserve the witness’s memory before it fades or is contaminated. Verbal overshadowing refers to research showing that describing something seen, and in particular describing a face, can under some conditions make that face harder to pick out afterward. The finding is counterintuitive because describing is exactly what a cooperative witness is asked to do, and it is one of the most closely examined effects in eyewitness research.
The effect reaches court in two ways. Defense counsel may argue that a detailed description, taken shortly before a lineup, weakened the witness’s memory for the culprit; prosecutors may respond that the effect is small and does not apply to the circumstances of the case. Courts also weigh the accuracy of the witness’s prior description as a factor in deciding whether an identification is reliable, and the research on descriptions bears on what that factor can show.
This paper summarizes the original studies, the meta-analysis and the large Registered Replication Report that measured the effect’s size and timing, current research on whether describing changes memory or decision-making, the relationship between description quality and identification accuracy, and the implications for how police take descriptions. It then reviews the federal, military, and Arizona law that governs prior descriptions and expert testimony on this subject.
The Original Finding
In the study that introduced the term, participants watched a brief videotape of a staged bank robbery (Schooler & Engstler-Schooler, 1990). One group spent several minutes writing out a detailed account of the robber’s face; a control group spent the same time on an unrelated activity. When both groups then viewed a photo lineup, the describers chose the robber less often than the controls. The authors named the effect verbal overshadowing and proposed that a verbal record of the face competed with, or overshadowed, the visual memory. The original experiments extended the finding to memory for colors and suggested that the effect is strongest for stimuli that are hard to put into words.
The original effect was large, and later research found it to be less consistent than the first reports suggested. Some studies replicated it, others did not, and a few found that describing improved later recognition. That uneven record led to two major attempts to measure the effect across many studies.
How Large Is the Effect, and When Does It Appear?
A meta-analysis of 29 effect-size comparisons, with 2,018 participants, found a small but statistically significant negative effect of describing on later identification (Meissner & Brigham, 2001). The effect depended on conditions: it was more likely when the identification test followed the description immediately, and more likely when participants were given elaborative description instructions rather than standard free-recall instructions. How a description is elicited also matters. In one series of experiments, the researchers varied how strict a standard participants used when reporting features; instructions to report every feature, even when unsure, produced the most incorrect descriptors and the poorest later identification, whereas instructions to report only what participants were confident of produced the fewest errors and the best identification performance among the description conditions (Meissner et al., 2001).
In 2014, 31 laboratories took part in a preregistered Registered Replication Report of the original bank-robbery study, among the first multi-laboratory Registered Replication Reports published by Perspectives on Psychological Science. The project ran two versions that differed only in the order of the description and a filler task. In the first, participants described the robber immediately after the video, then completed a 20-minute filler task before viewing a lineup. In the second, carried out in a smaller set of the participating laboratories, participants completed the 20-minute filler task first, then described the robber, then viewed the lineup. The effect was larger when the description was given after a delay following the event and the identification test followed soon after the description. Participants who described the robber were about 4 percentage points less likely than controls to identify him in the first version (about 51% of describers and 55% of controls) and about 16 percentage points less likely in the second (about 38% and 54%), with the sample sizes not stated here (Alogna et al., 2014). The first version, with an immediate description and a 20-minute delay before the lineup, produced the smaller effect. Both effects were smaller than the effect in the original study, but the second was robust across laboratories, and the authors concluded that the effect is real and strongly influenced by the timing of the tasks.
The timing pattern matters for application. A description given minutes after the event, with a substantial delay before any lineup, produced only a small effect in the replication; a description given after some time had passed, followed promptly by a lineup, produced a larger one. In practice, intervals are usually far longer than the minutes studied in the laboratory, and the research has not established how the effect behaves over the hours, days, or weeks that commonly separate a crime, a description, and a lineup.
Memory Impairment or a More Cautious Witness?
The replication measured only how often participants chose the culprit from a lineup that contained him. A drop in correct identifications can mean that describing impaired the witness’s memory, or that describing made the witness more cautious about choosing anyone. The two explanations have different legal implications. An impaired memory makes identifications less informative; a more cautious witness makes fewer identifications, but the identifications the witness does make may be as informative as before, or more so.
Several lines of research address the question. Studies in the 2000s found that describing could shift the witness’s decision criterion, making witnesses less willing to choose anyone from a lineup (Clare & Lewandowsky, 2004), and proposed that describing may change how witnesses search memory at test rather than altering the stored memory itself (Meissner et al., 2001). Commenting on the replication, Mickes and Wixted (2015) argued that a lower correct identification rate cannot by itself show impaired memory, because target-absent lineups are needed to measure false identifications as well; they cautioned that psychologists were not yet in a position to tell courts whether identifications made after a description deserve more or less weight.
More recent studies have used target-absent lineups and receiver operating characteristic (ROC) analysis, which measures how well witnesses separate guilty from innocent suspects. In four experiments with lineups and showups, Wilson, Seale-Carlisle, and Mickes (2018) found that describing the culprit immediately after the event did not reduce discriminability, but describing after a 20-minute delay did; a fifth experiment suggested that delayed descriptions contained a lower proportion of distinctive, diagnostic facial features. In a large online study that used the timing associated with the largest effect, describing the culprit reduced both correct identifications and discriminability, made witnesses more likely to say the culprit was not present, and was not prevented by a structured description method (the Person-Description Interview); in that study, the accuracy of the witnesses who did choose someone did not differ significantly by condition (Dodson et al., 2025). Using a formal model of lineup decisions, Therre et al. (2026) concluded that describing a culprit both reduced the likelihood that witnesses detected the culprit’s presence and reduced guessing-based choices. Taken together, these studies suggest that, under the conditions tested, describing can produce a modest loss of memory-based discrimination together with more cautious choosing, and that the second change may partly offset the first for witnesses who do make a choice.
Descriptions, Identification Accuracy, and Verbal Facilitation
Courts often treat an accurate description as a sign that a later identification is reliable. A meta-analysis of 33 research papers with 4,278 participants found a small but statistically significant relationship between how accurate a witness’s description was and whether the witness identified the right person (a correlation of about .14), and no meaningful positive relationship between the number of features described and identification accuracy (Meissner et al., 2008). The relationship was stronger in face-recognition paradigms than in eyewitness-identification paradigms. Lane and Houston (2021, pp. 39–42) summarize this literature as showing that description quality is a weak predictor of identification accuracy and that verbal overshadowing is small and inconsistent, and they note that describing a face sometimes improves later identification, a pattern researchers call verbal facilitation.
Description and identification also draw on different abilities. Most real-world descriptions are brief and general, focusing on sex, age, height, build, and clothing rather than inner facial features (Lane & Houston, 2021, pp. 28–29). A witness may recognize a face well while being unable to describe it, or describe a face fluently while recognizing it poorly. Neither a sparse description nor a detailed one, by itself, establishes whether a later identification is accurate.
Relevance to Police Description-Taking
None of this research suggests that police should stop taking descriptions. Descriptions are essential to an investigation, they create a record of the witness’s memory before later contamination, and the Supreme Court’s reliability framework treats the description as evidence in its own right. The research instead bears on how and when descriptions are taken and on how an identification that follows a description is evaluated.
Several practical points follow from the research, kept within its scope. An initial free-recall description, recorded verbatim in the witness’s own words, preserves the most useful evidence. Urging a witness to list every facial feature, to guess at features the witness is unsure of, or to describe the face again and again may increase the risk the laboratory studies identify, particularly when a lineup follows soon afterward. Building a feature-based facial composite involves similar analytic processing and is discussed in the companion paper on Face Composites. Recording the times of the event, the description, and any identification procedure allows the sequence to be evaluated later. These are inferences from laboratory work rather than field-tested procedures; the research has not established a specific interval or description method that eliminates the effect.
Limits of the Evidence
Most verbal overshadowing studies use students or online participants who watch a short video of a staged crime, describe a stranger for a few minutes, and view a photo lineup within half an hour, with no consequences for their choices. Real witnesses are often under stress, may have seen the culprit for a longer or shorter time, give descriptions in conversation with officers rather than in writing, and view lineups hours, days, or weeks later. The replication measured the effect at two specific timings within about 30 minutes; it did not test the intervals typical of real investigations. Effect sizes are reported as differences in group averages, and a modest average effect can coexist with no effect for many individuals and a larger effect for some; one large study reported that the effect varied with participants’ face-recognition ability (Dodson et al., 2025). No study can say whether a particular witness’s description affected that witness’s identification.
Questions to Ask About This Research
When an expert cites the Registered Replication Report (Alogna et al., 2014), counsel should ask which version of the procedure is being cited, how the timing of the description and lineup in that version compares with the timing in this case, and what is known about intervals longer than the 20 minutes studied. When a drop in correct identifications is offered as evidence of impaired memory, counsel should ask whether the study included culprit-absent lineups and measured discriminability, or whether the drop could reflect more cautious choosing (Mickes & Wixted, 2015; Wilson et al., 2018).
When the effect is described as small and therefore irrelevant, counsel should ask whether that conclusion follows: a small average effect neither proves that a description harmed a particular identification nor rules it out. When a witness’s accurate or detailed description is offered as proof that the identification is accurate, counsel should ask how strongly description accuracy predicts identification accuracy in the research (Meissner et al., 2008). In every case, counsel should ask what the record shows about how the description was elicited, how detailed it was, and how much time separated the event, the description, and the identification. These questions are explained in “Reading Research Claims: A Guide for Attorneys” on this site.
Verbal Overshadowing and the Law
Federal and Military Standards
The research bears most directly on the due process reliability analysis. Under Neil v. Biggers (1972) and Manson v. Brathwaite (1977), an identification obtained through an unnecessarily suggestive police procedure may still be admitted if it is reliable under the totality of the circumstances, considering the witness’s opportunity to view, degree of attention, accuracy of the prior description of the criminal, level of certainty at the confrontation, and the time between the crime and the confrontation. The description factor assumes that an accurate description signals a reliable identification; the research finds that relationship real but weak. Perry v. New Hampshire (2012) limited the due process screen to suggestive circumstances arranged by law enforcement. Taking a description is not a suggestive identification procedure, so arguments about verbal overshadowing ordinarily go to the weight of an identification and are presented through cross-examination, expert testimony, and argument rather than through suppression.
Expert testimony on the effect is governed by Federal Rule of Evidence 702, as amended December 1, 2023 (the proponent must show that it is more likely than not that each requirement is met, including a reliable application to the case), Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993), and Kumho Tire Co. v. Carmichael (1999), which extended the gatekeeping obligation to all expert testimony. In courts-martial, Military Rule of Evidence 321 governs eyewitness identification: an identification process is unreliable, and therefore unlawful, if it is so suggestive as to create a substantial likelihood of misidentification. Expert testimony is assessed under Military Rule of Evidence 702 and United States v. Houser (1993); under Military Rule of Evidence 1102, the 2023 federal amendment to Rule 702 became part of the military rule in 2025 (Joint Service Committee on Military Justice, 2025), and practitioners should confirm current text.
Arizona
Arizona challenges to pretrial identifications are heard at a Dessureault hearing outside the jury’s presence (State v. Dessureault, 1969), and Arizona courts assess the reliability of an identification that follows a suggestive procedure using the Biggers factors, including the accuracy of the prior description, which are not exclusive (State v. Rojo-Valenzuela, 2015). Arizona Rule of Evidence 702 adopted the Daubert standard effective January 1, 2012, and was amended effective January 1, 2024, to track the federal amendment (Ariz. Sup. Ct. Order R-23-0004). State v. Chapple (1983), Arizona’s leading case on eyewitness expert testimony, held that excluding such testimony was an abuse of discretion on its facts; outside circumstances like those, admission remains within the trial court’s discretion (State v. Nordstrom, 2001). Such testimony stays general, Arizona bars expert opinion on whether a particular witness is truthful (State v. Lindsey, 1986, 149 Ariz. at 475), and the expert’s role is to explain the factors rather than to say whether a particular identification is accurate.
Special Issues in Legal Practice
The principles described above apply with equal force to every party in a legal proceeding. The science operates the same way regardless of who is asking the questions, and the research summarized here does not by itself favor either side. What differs between the parties is not the science but the vantage point — the particular places where these issues tend to surface, and the particular errors each party is best positioned to guard against. The two summaries below identify those recurring issues. They are offered for orientation, not as litigation strategy; in every case the controlling question is what the evidence in that case will, and will not, support, and the expert’s role is to inform the trier of fact rather than to advocate for either side.
Special Issues for Prosecutors
The research supports several propositions prosecutors can legitimately present. Verbal overshadowing is a modest effect that varies with the timing of the description and the test, it was small in the replication when the description came immediately after the event and the lineup followed after a delay, and in some studies describing has no effect or improves recognition. Describing may make witnesses more cautious, and some studies found that the accuracy of witnesses who did choose was not significantly reduced. Taking a description is essential investigative practice, not a flaw. Prosecutors are well positioned to show, through a documented timeline and a verbatim record of the description, how the description was elicited and how much time passed before the identification.
The research also counsels caution. Under the conditions most favorable to the effect (about a 16-percentage-point drop in the second replication version), describing reduced correct identifications and, in recent ROC studies, reduced witnesses’ ability to discriminate guilty from innocent suspects; averaged across conditions, the effect is modest. A detailed or accurate description is only weakly related to identification accuracy and does not by itself corroborate an identification. Testimony or argument that verbal overshadowing has been disproved, or that it cannot apply outside the laboratory, goes beyond what the research shows.
Special Issues for Defense
The research supports several propositions defense counsel can legitimately present. The replication found a reliable effect, larger when a description was given after a delay and followed soon by the identification, and recent studies indicate that describing can reduce memory-based discrimination as well as willingness to choose. Pressing witnesses to report features they are unsure of impaired identification in experiments, and the accuracy of a description is a weak indicator of whether a later identification is correct. Expert testimony explaining these findings, tied to the documented sequence of description and identification, can help the factfinder weigh an identification.
The research also counsels caution about its limits. The effect is modest on average, depends on timing studied only over intervals of minutes, and in some conditions disappears or reverses. Its existence does not show that a described-then-identified face was misremembered, and an expert cannot say whether a particular description affected a particular witness. Descriptions remain essential to investigations, and treating the act of taking one as inherently corrupting misstates the research.
This paper is for educational purposes only and is not legal advice. The author is a psychologist, not an attorney; summaries of law are general and must be verified against current authority or with a licensed attorney.
References
Alogna, V. K., Attaya, M. K., Aucoin, P., Bahník, Š., Birch, S., Birt, A. R., Bornstein, B. H., Bouwmeester, S., Brandimonte, M. A., Brown, C., Buswell, K., Carlson, C., Carlson, M., Chu, S., Cislak, A., Colarusso, M., Colloff, M. F., Dellapaolera, K. S., Delvenne, J.-F., … Zwaan, R. A. (2014). Registered replication report: Schooler and Engstler-Schooler (1990). Perspectives on Psychological Science, 9(5), 556–578. https://doi.org/10.1177/1745691614545653
Clare, J., & Lewandowsky, S. (2004). Verbalizing facial memory: Criterion effects in verbal overshadowing. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30(4), 739–755. https://doi.org/10.1037/0278-7393.30.4.739
Dodson, C. S., Seale-Carlisle, T. M., Garrett, B. L., Kafadar, K., & Yaffe, J. (2025). Persistence of the verbal overshadowing and weapon-focus effects on lineup identification performance. Journal of Applied Research in Memory and Cognition, 14(3), 416–427. https://doi.org/10.1037/mac0000195
Joint Service Committee on Military Justice. (2025). Amendments to the Military Rules of Evidence since 1984 (G. E. Maggs, Ed.; 2025 ed.). https://jsc.defense.gov/Portals/99/MRE%20Amendments%202025.pdf
Lane, S. M., & Houston, K. A. (2021). Understanding eyewitness memory: Theory and applications. New York University Press.
Meissner, C. A., & Brigham, J. C. (2001). A meta-analysis of the verbal overshadowing effect in face identification. Applied Cognitive Psychology, 15(6), 603–616. https://doi.org/10.1002/acp.728
Meissner, C. A., Brigham, J. C., & Kelley, C. M. (2001). The influence of retrieval processes in verbal overshadowing. Memory & Cognition, 29(1), 176–186. https://doi.org/10.3758/BF03195751
Meissner, C. A., Sporer, S. L., & Susa, K. J. (2008). A theoretical review and meta-analysis of the description-identification relationship in memory for faces. European Journal of Cognitive Psychology, 20(3), 414–455. https://doi.org/10.1080/09541440701728581
Mickes, L., & Wixted, J. T. (2015). On the applied implications of the “verbal overshadowing effect.” Perspectives on Psychological Science, 10(3), 400–403. https://doi.org/10.1177/1745691615576762
Schooler, J. W., & Engstler-Schooler, T. Y. (1990). Verbal overshadowing of visual memories: Some things are better left unsaid. Cognitive Psychology, 22(1), 36–71. https://doi.org/10.1016/0010-0285(90)90003-M
Therre, A., Bell, R., Menne, N. M., Mayer, C., Lichtenhagen, U., & Buchner, A. (2026). On the cognitive processes underlying the verbal overshadowing effect: Culprit descriptions reduce culprit-presence detection and guessing-based selection in eyewitness responses to lineups. Legal and Criminological Psychology, 31(2), 301–315. https://doi.org/10.1111/lcrp.70013
Wilson, B. M., Seale-Carlisle, T. M., & Mickes, L. (2018). The effects of verbal descriptions on performance in lineups and showups. Journal of Experimental Psychology: General, 147(1), 113–124. https://doi.org/10.1037/xge0000354
Legal Authorities
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999).
Manson v. Brathwaite, 432 U.S. 98 (1977).
Neil v. Biggers, 409 U.S. 188 (1972).
Perry v. New Hampshire, 565 U.S. 228 (2012).
State v. Chapple, 135 Ariz. 281, 660 P.2d 1208 (1983).
State v. Dessureault, 104 Ariz. 380, 453 P.2d 951 (1969).
State v. Lindsey, 149 Ariz. 472, 720 P.2d 73 (1986).
State v. Nordstrom, 200 Ariz. 229, 25 P.3d 717 (2001).
State v. Rojo-Valenzuela, 237 Ariz. 448, 352 P.3d 917 (2015).
United States v. Houser, 36 M.J. 392 (C.M.A. 1993).
Fed. R. Evid. 702.
Ariz. R. Evid. 702.
Mil. R. Evid. 321.
Mil. R. Evid. 702.
Mil. R. Evid. 1102.
Related topic papers: Face Composites · Eyewitness Memory · Memory for Persons and Faces · Face Recognition · Improving Eyewitness Memory · Memory Expert Testimony
Related topics from Arizona Forensic Services: Courts-Martial