Snap Shot
- The question: Can a witness mistake a person who is familiar from an innocent encounter, such as a bystander at the scene, a customer, or a face seen in photographs, for the person who committed the crime?
- What the research shows: Laboratory studies show that prior exposure to an innocent person can raise the chance that a witness picks that person from a lineup, sometimes with high confidence. Meta-analytic evidence indicates the effect is larger after exposure to mugshots than after seeing a bystander at an event, and a recent large study shows it shrinks when the bystander looks different from the perpetrator or the perpetrator was seen longer.
- What it does not show: The effect is inconsistent and often small, several studies have not found it, and familiarity also supports accurate identification of people a witness truly knows. Research cannot say whether a particular witness’s identification of a particular familiar person is a transference error.
- The law: Mil. R. Evid. 321, the due process cases, and Arizona’s Dessureault hearing and RAJI Standard Criminal 45 govern identifications. Expert testimony is governed by Rule 702, Houser, Chapple, and Nordstrom, and in United States v. Garcia (1996) the Court of Appeals for the Armed Forces held harmless any error in excluding a defense expert’s testimony on memory transference.
- Issues for each side: Prosecutors should document every prior contact between the witness and the suspect and every photograph the witness viewed, without treating a familiar face as self-corroborating. Defense counsel should tie a transference claim to a documented innocent exposure and recognize that genuine acquaintance usually supports, rather than undermines, an identification.
A witness who identifies the wrong person does not always choose at random. Sometimes the person chosen really is familiar, but from somewhere other than the crime. Unconscious transference is the term for identifying an innocent but familiar person as the perpetrator: someone the witness saw at or near the scene, a regular customer, a neighbor, or a face the witness studied in a set of police photographs. The feeling that the face is known is genuine. The error lies in attributing that feeling to the crime. Because the recognition is real, the witness may be sincere and confident, and the mistake is not a lie.
The issue reaches court in two ways. A defendant may argue that the witness knew his or her face from an innocent encounter, such as a shop, a workplace, or an earlier photo array, and that the familiarity was mistaken for memory of the offense. A prosecutor may answer that the witness knew the defendant well and that familiarity is the reason the identification is reliable. Both claims have research support, and which applies depends on the kind and extent of the prior exposure.
This paper reviews the classic demonstrations, the meta-analytic evidence, failed replications and boundary conditions, competing explanations, and recent research, then the limits of the evidence and the military and Arizona law.
The Classic Demonstrations
Loftus (1976) drew early attention to the phenomenon in eyewitness identification. The best-known laboratory demonstration came from Ross et al. (1994), who in four experiments showed college students a film in which a male teacher read to preschool children and, in another scene, a different man stole money from a wallet. When shown a lineup that contained the teacher but not the thief, participants who had seen the teacher were nearly three times as likely as controls to choose him (p. 923). Most participants who made this error remembered the bystander vividly but believed that he and the thief were the same person, and the error was eliminated when participants were told before the lineup that the two were different people (pp. 923–926). Ross and colleagues therefore treated the error as resting on a conscious inference about identity rather than on familiarity alone (pp. 927–928; see also Earles et al., 2008).
Phillips et al. (1997) tested that interpretation with 650 undergraduates who viewed a videotaped robbery. Participants were more likely to choose the bystander than the actual perpetrator, and they were more confident when they chose the bystander. The misidentifications disappeared when the video was restaged so that the bystander and the perpetrator appeared together on screen for a few seconds. The authors concluded that a bystander’s presence alone was not enough, and that the error required the witness to infer that the two were one person.
Meta-Analytic Evidence: Mugshots and Bystanders
The most comprehensive quantitative review remains Deffenbacher et al. (2006). Across 32 independent tests, prior exposure to mugshots lowered the proportion of correct lineup decisions and raised the rate of false identifications, with the larger effect on false identifications. Whether the witness had identified someone from the mugshots, a “mugshot commitment” effect, was a substantial moderator of both effects. Simple retroactive interference, in which the perpetrator was not among the mugshots and no one from the mugshots reappeared in the lineup, did not significantly impair identification of the perpetrator. A third meta-analysis in the same review (Deffenbacher et al., 2006), of 19 independent tests of transference errors, found an effect more than twice as large in studies in which mugshot exposure occurred close in time to the target than in “bystander” studies with no mugshot exposure. A staged-crime study made the same distinction: witnesses who had picked someone from mug shots made more false lineup identifications than control witnesses, while witnesses who had merely seen the innocent person’s photograph did not differ from controls (Dysart et al., 2001).
These findings counsel precision. The term “unconscious transference” covers both the bystander seen at the event and the face seen in police photographs, and the evidence is considerably stronger for the second. Surveyed eyewitness experts have generally regarded the phenomenon as reliable enough to present in court; in a 2001 survey, unconscious transference was among the phenomena endorsed by at least 80% of respondents (Kassin et al., 2001); that survey predated much of the research reviewed here and measured expert opinion, not the size of the effect.
Failed Replications and Boundary Conditions
The bystander effect has not been found consistently. In five field studies with 330 retail store clerks and 340 students, Read et al. (1990) varied retention intervals, the time between seeing the bystander and the perpetrator, and lineup type and similarity. Witnesses who had seen the bystander generally did not misidentify the bystander more often than controls, and prior exposure often reduced misidentifications. The effect appeared only in the final experiment, under one combination of similarity and lineup construction that the authors judged unlikely in real settings. A familiar face in the lineup did make witnesses more likely to choose someone, and the authors suggested that misidentifications may rest partly on inferences that the bystander could be the perpetrator. Deffenbacher and colleagues’ (2006) finding that pure bystander studies produced smaller effects than mugshot-based studies is consistent with that record.
Research since then has identified conditions that make the error more or less likely. A 2024 experiment with 31,612 participants varied how strongly the perpetrator and the bystander were encoded and how similar they looked. Transference occurred, but it was much weaker when the bystander looked different from the perpetrator and when the perpetrator had been viewed longer than the bystander (Carlson et al., 2024). Witness age also matters. Among children aged 5 to 12, only 11- and 12-year-olds were more likely than controls to choose a familiar bystander (Ross et al., 2006), and in a study of 431 participants, adolescents aged 14 to 16 were more prone to choose the bystander than younger children or adults (Brackmann et al., 2019). Older adults made more transference errors than younger adults when the faces were young, but not when the faces were of their own age group (Perfect & Harris, 2003). In a small Spanish study of 57 adults with intellectual disabilities, the bystander was chosen more often than other lineup members whether or not the perpetrator was present (Vilariño et al., 2025).
Why It Happens: Competing Explanations
The most widely used account is source monitoring. Memories carry no label stating where they came from; people infer the source from the qualities of the memory, such as perceptual detail and context, and errors are more likely when the sources are similar or when encoding was limited (Johnson et al., 1993). Under this view, the witness correctly senses that a face is familiar but attributes the familiarity to the wrong occasion. Dual-process theories make a related point: a sense of familiarity without recollection of where and when a face was seen is an unsafe basis for an identification (Lane & Houston, 2021, pp. 48–51).
A second account, memory blending, holds that the memories of the two people are merged, so that the witness comes to remember them as one person; it is related to the identity inference that Ross et al. (1994) described. A third, related account treats some transference errors as a form of change blindness. In three experiments using a filmed supermarket scene, an innocent person walked down an aisle and passed behind boxes, and the thief then emerged, creating an illusion of continuity. More than half the participants did not notice the change of person, and those who missed it were more likely to choose the innocent person who appeared continuous with the thief than one shown elsewhere in the store; distraction made the change harder to notice (Davis et al., 2008).
The accounts are not mutually exclusive. Perfect and Harris (2003) found evidence of both source errors and identity confusions. In two experiments, viewers more often credited a person with another person’s actions when the two had appeared in the same event, and those errors came with high confidence and reports of vividly “remembering” (Kersten et al., 2013). The large 2024 study found the most support for familiarity-based and source-retrieval explanations and the least for memory blending (Carlson et al., 2024). In practical terms, a transference error is most plausible when a witness had a real basis for familiarity, weak memory of the perpetrator, and some reason to link the two people in time or place.
Recent Research, 2022–2026
Two recent studies from one laboratory add to the picture. The 2024 study described above also found that transference weakened the confidence–accuracy relationship (Carlson et al., 2024). In two 2026 experiments, participants studied same- and other-race faces, then viewed a perpetrator and a lineup whose suspect was guilty or innocent and either familiar or unfamiliar. Familiar innocent suspects of either race were vulnerable to misidentification, the usual cross-race effect appeared only for unfamiliar suspects, and adding conceptual information about the familiar faces weakened the confidence–accuracy relationship (Wooten et al., 2026). The authors concluded that casual familiarity can create risk rather than reduce it. Both studies created familiarity through brief exposures, far weaker than knowing someone personally.
Familiarity Can Also Support Accurate Identification
The same research that documents transference shows that genuine familiarity is a strength of face memory. People recognize faces they know well from poor images and across changes in lighting and angle, while recognition of strangers is much less reliable. In one study, viewers who knew the people in poor-quality security video identified them with high accuracy, while viewers who did not, including experienced police officers, performed poorly (Burton et al., 1999). Most sexual assaults reported by women involve a perpetrator the complainant knew (Lane & Houston, 2021, pp. 44, 56), and in such cases identity is usually not a recognition problem.
The difficulty lies between the extremes. Familiarity is a matter of degree, and research has not established how accurately witnesses identify people they know only slightly or have seen only a few times (Lane & Houston, 2021, pp. 54–57). A 2019 review devoted to identifications by eyewitnesses who were familiar with the person identified surveyed what is known in this area (Vallano et al., 2019). The transference research applies most directly to that middle ground: a face the witness has seen, but whose source the witness may not be able to place.
Limits of the Evidence
Nearly all of this research uses staged events, filmed thefts, or photographs, with student or online participants who have no stake in the outcome. Exposures to the bystander are brief and controlled, and many lineups omit the perpetrator so that a bystander choice can be counted as an error. Real cases differ: witnesses may have seen an innocent person many times, under unrecorded conditions, and the perpetrator may or may not be in the lineup. Effect sizes vary widely across studies, the bystander effect in particular has been inconsistent, and the strongest evidence concerns mugshot exposure rather than chance encounters at a scene. Most important, group research shows how often familiar innocent people are chosen under certain conditions. It cannot show whether a particular witness’s identification of a particular familiar person is accurate or a transference error.
Questions to Ask About This Research
When Ross et al. (1994) or Phillips et al. (1997) is cited, counsel should ask whether this witness had any basis to tell the two people apart, because in those studies the effect disappeared when witnesses were told, or saw, that they were different people. When Deffenbacher et al. (2006) is cited, counsel should ask which of its findings applies, because the effect was large after mugshot commitment and much smaller in pure bystander studies. When an expert relies on Carlson et al. (2024), counsel should ask how similar the familiar person is to the description of the perpetrator and how long the witness viewed each person, because those factors substantially changed the size of the effect.
Questions about the case facts matter as much. What is the evidence that the witness saw the defendant before the crime, where, how often, and under what conditions? Did the witness view the defendant’s photograph in any earlier procedure, and did the witness choose anyone at that time? Was the witness asked at the identification whether, and from where, the person chosen was known? Does the expert claim that the research can determine whether this identification was a transference error, and if so, on what study? These questions are explained in “Reading Research Claims: A Guide for Attorneys” on this site.
Unconscious Transference and the Law
Federal and Military Standards
The due process screen for identifications asks whether a police-arranged procedure was unnecessarily suggestive and, if so, whether the identification is nonetheless reliable under the factors of Neil v. Biggers (1972) and Manson v. Brathwaite (1977). Familiarity from an innocent encounter that police did not arrange does not trigger that screen; in Perry v. New Hampshire (2012), the Supreme Court left the reliability of such identifications to cross-examination, jury instructions, and expert testimony. Familiarity created by an earlier police procedure, such as a mugshot search or a prior array containing the suspect, is part of the procedural history a court examines. In courts-martial, Mil. R. Evid. 321 makes an identification inadmissible if it results from a lineup or other identification process so suggestive as to create a substantial likelihood of misidentification, or if due process requires exclusion (Mil. R. Evid. 321(b)–(c)). When the defense raises unreliability, the prosecution must show by a preponderance of the evidence that the identification was reliable under the circumstances, and a later identification after an unreliable one is admissible only on clear and convincing evidence that it was not the result of the inadmissible one (Mil. R. Evid. 321(d)(6)(B)).
Expert testimony on eyewitness memory is governed by Mil. R. Evid. 702 and the six factors of United States v. Houser (1993). In United States v. Garcia (1996), the military judge excluded a defense eyewitness expert’s testimony on the confidence–accuracy relationship and on theories of memory transference and transposition. The Air Force Court of Military Review found an abuse of discretion but no prejudice, and the Court of Appeals for the Armed Forces affirmed, holding that any error was harmless beyond a reasonable doubt; it noted that the transference testimony was theoretical and general and added little to a police officer’s concessions about the suggestive procedures (44 M.J. at 28, 31, 33). The December 1, 2023 amendment to Fed. R. Evid. 702, which requires the proponent to show that it is more likely than not that each admissibility requirement is met, became part of Mil. R. Evid. 702 in 2025 through Mil. R. Evid. 1102 (Joint Service Committee on Military Justice, 2025); practitioners should confirm the current text. Under either rule, an expert may explain the conditions under which transference occurs but may not opine that a particular identification is mistaken.
Arizona
When a defendant challenges a pretrial identification, an Arizona trial court holds a hearing outside the jury’s presence under State v. Dessureault (1969); if the pretrial procedure was unduly suggestive, the State must show by clear and convincing evidence that the in-court identification is not tainted by it. Arizona’s standard instruction, RAJI (Criminal) Standard Criminal 45, tells jurors that the State must prove beyond a reasonable doubt that the in-court identification is reliable and lists factors drawn from Biggers and Manson, ending with “any other factor that affects the reliability of the identification,” which allows argument about a witness’s prior exposure to the defendant. State v. Chapple (1983) is Arizona’s leading case on eyewitness expert testimony: the Arizona Supreme Court held that excluding such testimony was an abuse of discretion in that case, where identification was the central disputed issue, and admission otherwise remains discretionary. State v. Nordstrom (2001) approved allowing an expert to explain factors affecting reliability while barring opinions on a particular witness’s accuracy. Arizona Rule of Evidence 702 adopted the Daubert standard effective January 1, 2012 and was amended effective January 1, 2024 to track the 2023 federal amendment (Ariz. Sup. Ct. Order R-23-0004), and State v. Lindsey (1986) bars expert opinions on whether a particular witness is telling the truth.
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. The bystander effect is inconsistent and often small, several studies have not found it, and it weakens substantially when the familiar person looks different from the perpetrator or when the witness had a longer view of the perpetrator. Identification of a person the witness genuinely knows well rests on one of the strongest capacities of face memory, and many cases involve an acquaintance whose identity is not seriously in doubt. A transference argument is much stronger when there is evidence of an actual innocent exposure, and weak when it rests only on the possibility of one. Prosecutors are well positioned to make an identification stronger by recording every prior contact between the witness and the suspect, every photograph the witness viewed and any choice the witness made, and the witness’s own statement, at the first identification, of whether and from where the person chosen was known.
The research also counsels caution. Exposure to a suspect’s photograph in an earlier procedure, and especially an earlier choice of that person, can increase false identifications, and a later lineup that repeats the suspect cannot be treated as an independent test. A witness’s confidence is less informative when the person chosen was familiar from another source, so a confident identification of a casually familiar suspect is not self-corroborating. Familiarity with a suspect who is only slightly known, such as a customer or a face from the neighborhood, falls in the range the research has examined least, and testimony that treats any familiarity as a guarantee of accuracy overstates the evidence.
Special Issues for Defense
The research supports several propositions defense counsel can legitimately present. A witness can mistake a person familiar from an innocent encounter for the perpetrator, sincerely and with confidence. The risk is greatest when the witness viewed the defendant’s photograph before the lineup, and especially when the witness chose that photograph; when the defendant and the perpetrator look alike or were seen close together in time or place; and when the witness’s view of the perpetrator was brief. Expert testimony describing these conditions, combined with evidence of the witness’s actual prior exposure to the defendant, is a legitimate way to help the factfinder evaluate the identification, and Chapple recognizes that such testimony can matter; in Garcia, by contrast, the court held any error in excluding the expert’s testimony harmless, noting that the general transference testimony added little where the suggestive procedures were already before the members.
The research also counsels caution about the limits that apply to the defense. The bystander effect has failed to replicate in several studies and is often small, so the mere possibility that a witness saw the defendant somewhere is weak support for a transference claim. Where the witness knew the defendant well, familiarity is more likely to support accuracy than to undermine it, and an expert who suggests otherwise goes beyond the research. An expert cannot tell the factfinder that a particular identification was a transference error; the research describes conditions and probabilities, and the application to the case belongs to the factfinder.
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
Brackmann, N., Sauerland, M., & Otgaar, H. (2019). Developmental trends in lineup performance: Adolescents are more prone to innocent bystander misidentifications than children and adults. Memory & Cognition, 47(3), 428–440. https://doi.org/10.3758/s13421-018-0877-6
Burton, A. M., Wilson, S., Cowan, M., & Bruce, V. (1999). Face recognition in poor-quality video: Evidence from security surveillance. Psychological Science, 10(3), 243–248. https://doi.org/10.1111/1467-9280.00144
Carlson, C. A., Wooten, A. R., & Carlson, M. A. (2024). Was he the perpetrator or a bystander? Testing theories of unconscious transference for eyewitness identification. Journal of Applied Research in Memory and Cognition, 13(4), 543–555. https://doi.org/10.1037/mac0000155
Davis, D., Loftus, E. F., Vanous, S., & Cucciare, M. (2008). “Unconscious transference” can be an instance of “change blindness.” Applied Cognitive Psychology, 22(5), 605–623. https://doi.org/10.1002/acp.1395
Deffenbacher, K. A., Bornstein, B. H., & Penrod, S. D. (2006). Mugshot exposure effects: Retroactive interference, mugshot commitment, source confusion, and unconscious transference. Law and Human Behavior, 30(3), 287–307. https://doi.org/10.1007/s10979-006-9008-1
Dysart, J. E., Lindsay, R. C. L., Hammond, R., & Dupuis, P. (2001). Mug shot exposure prior to lineup identification: Interference, transference, and commitment effects. Journal of Applied Psychology, 86(6), 1280–1284. https://doi.org/10.1037/0021-9010.86.6.1280
Earles, J. L., Kersten, A. W., Curtayne, E. S., & Perle, J. G. (2008). That’s the man who did it, or was it a woman? Actor similarity and binding errors in event memory. Psychonomic Bulletin & Review, 15(6), 1185–1189. https://doi.org/10.3758/PBR.15.6.1185
Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin, 114(1), 3–28. https://doi.org/10.1037/0033-2909.114.1.3
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
Kassin, S. M., Tubb, V. A., Hosch, H. M., & Memon, A. (2001). On the “general acceptance” of eyewitness testimony research: A new survey of the experts. American Psychologist, 56(5), 405–416. https://doi.org/10.1037/0003-066X.56.5.405
Kersten, A. W., Earles, J. L., & Upshaw, C. (2013). False recollection of the role played by an actor in an event. Memory & Cognition, 41(8), 1144–1158. https://doi.org/10.3758/s13421-013-0334-5
Lane, S. M., & Houston, K. A. (2021). Understanding eyewitness memory: Theory and applications. New York University Press.
Loftus, E. F. (1976). Unconscious transference in eyewitness identification. Law and Psychology Review, 2, 93–98.
Perfect, T. J., & Harris, L. J. (2003). Adult age differences in unconscious transference: Source confusion or identity blending? Memory & Cognition, 31(4), 570–580. https://doi.org/10.3758/BF03196098
Phillips, M. R., Geiselman, R. E., Haghighi, D., & Lin, C. (1997). Some boundary conditions for bystander misidentification. Criminal Justice and Behavior, 24(3), 370–390. https://doi.org/10.1177/0093854897024003004
Read, J. D., Tollestrup, P., Hammersley, R., McFadzen, E., & Christensen, A. (1990). The unconscious transference effect: Are innocent bystanders ever misidentified? Applied Cognitive Psychology, 4(1), 3–31. https://doi.org/10.1002/acp.2350040103
Ross, D. F., Ceci, S. J., Dunning, D., & Toglia, M. P. (1994). Unconscious transference and mistaken identity: When a witness misidentifies a familiar but innocent person. Journal of Applied Psychology, 79(6), 918–930. https://doi.org/10.1037/0021-9010.79.6.918
Ross, D. F., Marsil, D. F., Benton, T. R., Hoffman, R., Warren, A. R., Lindsay, R. C. L., & Metzger, R. (2006). Children’s susceptibility to misidentifying a familiar bystander from a lineup: When younger is better. Law and Human Behavior, 30(3), 249–257. https://doi.org/10.1007/s10979-006-9034-z
Vallano, J. P., Slapinski, K. A., Steele, L. J., Briggs, A. P., & Pozzulo, J. D. (2019). Familiar eyewitness identifications: The current state of affairs. Psychology, Public Policy, and Law, 25(3), 128–146. https://doi.org/10.1037/law0000204
Vilariño, M., Brito, D., Sanmarco, J., & González Amado, B. (2025). Performance in identification line-ups and unconscious transference in individuals with intellectual disabilities. Publicaciones, 55(2), 75–91. https://doi.org/10.30827/publicaciones.v55i2.35406
Wooten, A. R., Carlson, C. A., Munir, S., Carlson, M. A., & Olof, H. (2026). Investigating the impact of prior suspect familiarity on same- and cross-race lineup identifications. Applied Cognitive Psychology, 40(4), Article e70236. https://doi.org/10.1002/acp.70236
Legal Authorities
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).
United States v. Garcia, 44 M.J. 27 (C.A.A.F. 1996).
United States v. Houser, 36 M.J. 392 (C.M.A. 1993).
U.S. Const. amend. V.
U.S. Const. amend. XIV.
Ariz. Sup. Ct. Order R-23-0004 (2023) (amending Ariz. R. Evid. 702, effective Jan. 1, 2024).
Fed. R. Evid. 702.
Ariz. R. Evid. 702.
Rev. Ariz. Jury Instr. (Crim.) Standard Criminal 45 (Identification) (6th ed. 2022).
Mil. R. Evid. 321.
Mil. R. Evid. 702.
Mil. R. Evid. 1102.
Related topic papers: Mug Books and Mugshot Exposure · Source Monitoring · Change Blindness and Inattentional Blindness · Face Recognition · Lineup Construction · Confidence and Accuracy
Related topics from Arizona Forensic Services: Courts-Martial