Snap Shot
- The question: When an account of an event passes from one person to another, or is retold by the same person, how does it change, and what survives?
- What the research shows: Retold accounts generally become shorter (leveling), keep and emphasize a smaller set of details (sharpening), and drift toward what the teller and audience expect (assimilation). Retelling for a purpose or an audience can bias the teller’s own later memory. The gist and the most important or surprising elements are often well preserved, and familiar material and audiences that demand accuracy reduce change.
- What it does not show: The research does not show that any particular second-hand or retold account is inaccurate, and it cannot identify which details in an account were lost or added. Leveling–sharpening as a cognitive-style trait is a dated construct with inconsistent findings and no validated use for classifying a witness.
- The law: Expert testimony on these processes is governed by Rule 702, Houser in courts-martial, and Chapple and Nordstrom in Arizona, and stays general. Accounts that pass through several people reach court under the hearsay-within-hearsay rule (Fed., Mil., and Ariz. R. Evid. 805), under which each layer must separately qualify.
- Issues for each side: Prosecutors can show that core content is often preserved across retellings but should not treat agreement among second-hand reports as independent corroboration; defense counsel can show how detail is lost and reshaped along a chain but should tie the argument to the actual links and content.
Much of the information in a criminal case comes second-hand. A complainant tells a friend, who tells a parent, who calls the police; an officer summarizes a bystander’s statement; a command learns of an allegation through several layers of reporting; witnesses in a unit or a school hear one another’s versions before they are interviewed. A person may also tell the same event many times, to different listeners, for different purposes. Each telling is an occasion on which the account can be shortened, emphasized differently, or brought into line with what the teller or listener expects.
Leveling and sharpening are the classic terms for two of these changes. Leveling is the loss of detail as an account is retold, so that it becomes shorter. Sharpening is the selective retention, and often the emphasis, of the details that remain. Assimilation is the reshaping of what remains to fit expectations, interests, and beliefs. None of these involves lying; they are ordinary features of remembering and communicating.
This paper traces the concept from Bartlett and from Allport and Postman through the cognitive-style research of the 1950s and 1960s, summarizes modern research on serial reproduction and on retelling one’s own memories, and addresses second-hand accounts, chains of reporting, the limits of the evidence, and the governing law. Related processes are covered in Reconstruction and Recall, Schema and Script Memory, Remembering Changes Memory, and External Sources of Memory Distortion.
Origins: Bartlett and the Psychology of Rumor
The method goes back to Bartlett (1932), who asked people to reproduce an unfamiliar folk tale, “The War of the Ghosts,” either repeatedly over time or in a chain in which each person’s version became the next person’s material. The stories tended to become shorter, and unfamiliar elements were dropped or recast in more familiar terms. Bartlett’s view that remembering is an effort to make sense of material using prior knowledge became a foundation of the modern view of memory as reconstructive.
Later scholars have cautioned against reading Bartlett as having shown that memory is inherently unreliable; he also described detailed and accurate recall and sought to explain it within his theory (Ost & Costall, 2002). One of his best-known illustrations, a drawing that drifted toward a face as it was reproduced, did not replicate in five experiments (Carbon & Albrecht, 2012).
The terms leveling and sharpening were popularized by Allport and Postman (1945, 1947) in wartime research on rumor, in which a description of a picture was passed from person to person. They reported that accounts lost detail quickly over the first several retellings and then stabilized in a short form (leveling), that a few details were retained and became more prominent (sharpening), and that content was reshaped to fit participants’ interests and expectations, including stereotypes (assimilation).
That research has often been misdescribed. It is frequently cited as an eyewitness experiment showing that observers moved a razor from a White man’s hand to a Black man’s. Treadway and McCloskey (1987) documented that the eyewitness literature had misdescribed the procedure, a serial retelling of a description rather than a test of what eyewitnesses perceived or remembered, and had enlarged its results (see also Treadway & McCloskey, 1989). The study remains a foundation for research on rumor, but it is not evidence about how a single eyewitness perceives an event.
Leveling–Sharpening as a Cognitive Style
In the 1950s and 1960s, researchers at the Menninger Foundation used the same terms for a different idea: a stable individual difference in how memories of similar experiences blend together. In perceptual tasks involving a series of stimuli, people classified as “levelers” showed stronger assimilation of each impression to earlier ones, while “sharpeners” kept successive impressions more distinct (Holzman & Klein, 1954), and the two groups differed in their recall of a familiar children’s story (Holzman & Gardner, 1960).
This work did not develop into a validated tool. Scores on a standard leveling–sharpening task reflected, in part, how participants interpreted the instructions and their motivation (Gilgen & Bakan, 1968), and one study found no relation between leveling–sharpening and memory or a repressive coping style (Lewinsohn et al., 1970). The research is dated, its findings are inconsistent, and it provides no basis for classifying a witness as a “leveler” or predicting how an individual’s account will change.
Serial Reproduction: What Changes and What Survives
Modern transmission-chain studies, in which material is passed along a line of participants, have identified content biases: regular tendencies for some kinds of information to be passed on more faithfully than others (Mesoudi & Whiten, 2008). When descriptions of everyday events written as low-level actions were passed along chains of four people, low-level detail declined with each generation and the accounts were increasingly told in summary terms (Mesoudi & Whiten, 2004). That is leveling in measurable form.
Chains also show assimilation to shared expectations. When a story about a man and a woman was passed along five-person chains, early participants sometimes reproduced stereotype-inconsistent information more than consistent information, but by the end of the chains stereotype-consistent information was retained better (Kashima, 2000).
The same research shows what retelling preserves. In a study with 2,389 participants, retellers closely preserved how surprising a story was across generations of retelling, even in chains where the surprising event itself had been lost or altered (Breithaupt et al., 2018). A familiar urban legend was recalled better across five reproductions than “The War of the Ghosts,” and recall was better when participants expected a strict rather than a lenient audience (Ost et al., 2022). Across these studies, the gist and emotional character of a story tend to survive, while specific details, especially unexpected or peripheral ones, are the most likely to be lost or reshaped.
Retelling One’s Own Memory
People also level and sharpen their own experiences, and the retelling can affect what they later remember. In four experiments, participants who read a story and wrote a biased letter about one character later recalled the story in line with that perspective, in what they recalled and in the direction of their errors, even when the letters contained little specific information (Tversky & Marsh, 2000). In a diary study, participants labeled 61% of their everyday retellings as distorted by exaggeration, omission, minimization, or addition, but only 42% as inaccurate, and they exaggerated to entertain and simplified to inform (Marsh & Tversky, 2004). Conversational retelling serves the speaker’s goals and audience, not the complete and accurate recall that laboratory tests demand (Marsh, 2007).
The audience matters. A speaker who tailors a description of a person to a listener’s apparent views later tends to remember the person in line with the tailored description (Higgins & Rholes, 1978), especially when the speaker trusts the audience (Echterhoff et al., 2005); across three studies with 1,070 participants, the bias appeared whether or not the speaker had first formed an evaluative judgment of the person (Wagner & Echterhoff, 2025).
Retelling is not simply harmful. Participants who retold stories as if talking to friends later recalled more specific details than participants who did not retell them, although fewer than those who retold them as accurately as possible (Harris & McDermott, 2025). The research does not support a presumption that a frequently retold account is unreliable; it supports asking how, why, and to whom it was retold.
Chains of Reporting in Legal Cases
A parent describes what a child said at bedtime; a friend describes a disclosure at a party; a supervisor relays what one service member told another. Each link is a person who heard an account, understood it in a particular way, and retold it for a particular purpose. Listeners generally retain the meaning of what was said far better than its exact words (Davis & Friedman, 2007), so a relayed statement is ordinarily the relayer’s version of the meaning, not a transcript.
When investigators’ contemporaneous “verbatim” notes of 20 forensic interviews with children were compared with recordings, the notes omitted 57% of the interviewers’ utterances and 25% of the incident-relevant details the children provided, and fewer than half of the children’s details were linked to the type of question that actually produced them, with investigators tending to credit details to more open questions than had been asked (Lamb et al., 2000). Experienced interviewers recalled much of the substance of what children told them but often could not recall their own leading questions (Warren & Woodall, 1999). In both studies, notes and recollections frequently lost the context of a statement, including the question that elicited it, and in the Lamb et al. study the notes also omitted about one quarter of the incident-relevant details.
Talk among witnesses is a related chain, covered in External Sources of Memory Distortion. When several people report an account they all heard from one source, their agreement reflects the source, not independent confirmation, and the elements that circulate most stably tend to be the shared, expected, and striking ones, which may or may not matter most to the legal questions.
The research also supports reliance on chains of reporting for some purposes. The gist, emotional character, and central claim of a disclosure are the elements most likely to survive retelling, and an early, recorded, first-hand account is the best protection against later change.
Limits of the Evidence
Most serial reproduction and retelling studies use stories or pictures with no personal significance, usually retold in writing by university students within minutes or days. A disclosure of abuse or assault to a trusted person differs in emotional significance and in the listener’s motivation to remember accurately, which may increase or decrease change. Effects vary with the material, chain length, and instructions, and several classic demonstrations have been misdescribed or have not replicated. Group findings cannot show that a particular relayed or retold account was distorted, and without a record of each link an expert cannot determine which details changed.
Questions to Ask About This Research
When the Allport and Postman research is cited, counsel should ask what the study actually did, in light of Treadway and McCloskey (1987). When an expert describes leveling in a chain of reporting, counsel should ask how many links the chain contains, whether any link was recorded, what each relayer’s purpose was, and whether the cited studies used personally significant material and conversation rather than written reproduction of a story. When Lamb et al. (2000) is cited, counsel should ask whether the interview was recorded; the study shows that notes are a poor substitute for a recording, not that a recorded interview is inaccurate. When Harris and McDermott (2025) or Breithaupt et al. (2018) is offered to show that retelling preserves content, counsel should ask which content: those studies support preservation of gist and affect, not of every detail. These questions are explained in “Reading Research Claims: A Guide for Attorneys” on this site.
Leveling and Sharpening and the Law
Federal and Military Standards
Expert testimony about retelling is governed by Federal Rule of Evidence 702, as amended December 1, 2023, under which the proponent must show that it is more likely than not that the rule’s requirements are met, including a reliable application of reliable methods to the case. Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993) made the trial judge a gatekeeper, and Kumho Tire Co. v. Carmichael (1999) extended that obligation to all expert testimony, including technical or other specialized knowledge. Military Rule of Evidence 702 tracks the federal rule, and the 2023 amendment took effect in the military rule in 2025 through Military Rule of Evidence 1102 (Joint Service Committee on Military Justice, 2025); practitioners should confirm current text. Military judges apply the six factors of United States v. Houser (1993). Such testimony is ordinarily general and may not tell the factfinder that a particular account is accurate or false.
Under Federal Rule of Evidence 805, hearsay within hearsay is not excluded if each part of the combined statements conforms with an exception to the hearsay rule; Military Rule of Evidence 805 is the same except that it refers to an exception or exclusion. Each link in a chain must therefore be separately justified. Under Rule 806 in both systems, when a hearsay statement has been admitted, the declarant’s credibility may be attacked, and then supported, by evidence that would be admissible if the declarant had testified. When a link in the chain is a testimonial statement by a declarant who does not testify, the Confrontation Clause bars its use against the accused unless the declarant is unavailable and the accused had a prior opportunity for cross-examination (Crawford v. Washington, 2004). These rules decide admissibility; the research bears on the weight of a retold statement.
Arizona
Arizona Rule of Evidence 702 adopted the federal formulation and the Daubert standard effective January 1, 2012, and was amended effective January 1, 2024, to conform to the 2023 federal amendment (Ariz. Sup. Ct. Order R-23-0004). State v. Chapple (1983), Arizona’s leading case on memory and eyewitness expert testimony, held exclusion of a qualified expert an abuse of discretion on its facts, and State v. Nordstrom (2001) approved limiting such an expert to general factors without opinions on a particular witness’s accuracy. Experts may not opine on whether a particular witness is truthful (State v. Lindsey, 1986, 149 Ariz. at 475). Arizona Rules of Evidence 805 and 806 follow the federal rules; practitioners should confirm current text.
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. Retelling often, but not always, preserves the gist, central claim, and emotional character of an account, accuracy-focused retelling can strengthen memory for what is retold (slanted retelling, repeated suggestive questioning, or exposure to others’ accounts can contaminate it; see “Remembering Changes Memory”), and the fact that a complainant told several people does not by itself make the account unreliable. Prosecutors are well positioned to reduce the effects described here by obtaining early, first-hand, recorded statements, recording forensic interviews rather than relying on notes, and documenting who told whom what, and when.
The research also counsels caution. Agreement among witnesses who each heard the account from the same person is not independent corroboration. A relayer’s report is ordinarily the relayer’s version of the meaning, and specific details, wording, and the questions that prompted a statement are the elements most likely to be lost or reshaped. Presenting second-hand accounts as verbatim, or treating details added along a chain as confirming the original, goes beyond what the research supports.
Special Issues for Defense
The research supports several propositions defense counsel can legitimately present. Accounts passed from person to person generally lose detail, keep a selected subset of details, and drift toward what tellers and listeners expect, and retelling for a purpose or audience can bias the teller’s own later memory. In two older studies, investigators’ notes and recollections omitted many of the questions asked, and in one of them details tended to be credited to more open questions than were actually asked; whether that occurred in a particular case requires a recording or other record. Expert testimony on these principles, combined with evidence of the actual links in a chain and the content at each link, is a legitimate way to help the factfinder weigh a retold account.
The research also counsels caution about the limits that apply to the defense. Retelling does not by itself make an account inaccurate, and gist and central claims are often well preserved. An expert cannot say which details in a particular account were lost or added, or that a particular relayer distorted a statement. The Allport and Postman study is not an eyewitness experiment, and the cognitive-style research does not support labeling a witness a “leveler.” The same principles apply to relayed accounts of what the accused said.
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
Allport, G. W., & Postman, L. (1945). The basic psychology of rumor. Transactions of the New York Academy of Sciences, 8(2, Series II), 61–81. https://doi.org/10.1111/j.2164-0947.1945.tb00216.x
Allport, G. W., & Postman, L. (1947). The psychology of rumor. Henry Holt.
Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. Cambridge University Press.
Breithaupt, F., Li, B., Liddell, T. M., Schille-Hudson, E. B., & Whaley, S. (2018). Fact vs. affect in the telephone game: All levels of surprise are retold with high accuracy, even independently of facts. Frontiers in Psychology, 9, Article 2210. https://doi.org/10.3389/fpsyg.2018.02210
Carbon, C.-C., & Albrecht, S. (2012). Bartlett’s schema theory: The unreplicated “portrait d’homme” series from 1932. Quarterly Journal of Experimental Psychology, 65(11), 2258–2270. https://doi.org/10.1080/17470218.2012.696121
Davis, D., & Friedman, R. D. (2007). Memory for conversation: The orphan child of witness memory researchers. In M. P. Toglia, J. D. Read, D. F. Ross, & R. C. L. Lindsay (Eds.), The handbook of eyewitness psychology: Vol. 1. Memory for events (pp. 3–52). Lawrence Erlbaum Associates.
Echterhoff, G., Higgins, E. T., & Groll, S. (2005). Audience-tuning effects on memory: The role of shared reality. Journal of Personality and Social Psychology, 89(3), 257–276. https://doi.org/10.1037/0022-3514.89.3.257
Gilgen, A. R., & Bakan, P. (1968). Influence of task and subject variables in the Squares Test of leveling-sharpening. Perceptual and Motor Skills, 27(1), 211–218. https://doi.org/10.2466/pms.1968.27.1.211
Harris, K., & McDermott, K. (2025). The effect of social retelling on event recall. Memory, 33(2), 166–177. https://doi.org/10.1080/09658211.2024.2415907
Higgins, E. T., & Rholes, W. S. (1978). “Saying is believing”: Effects of message modification on memory and liking for the person described. Journal of Experimental Social Psychology, 14(4), 363–378. https://doi.org/10.1016/0022-1031(78)90032-X
Holzman, P. S., & Gardner, R. W. (1960). Leveling-sharpening and memory organization. Journal of Abnormal and Social Psychology, 61(2), 176–180. https://doi.org/10.1037/h0041295
Holzman, P. S., & Klein, G. S. (1954). Cognitive system-principles of leveling and sharpening: Individual differences in assimilation effects in visual time-error. Journal of Psychology, 37(1), 105–122. https://doi.org/10.1080/00223980.1954.9916135
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
Kashima, Y. (2000). Maintaining cultural stereotypes in the serial reproduction of narratives. Personality and Social Psychology Bulletin, 26(5), 594–604. https://doi.org/10.1177/0146167200267007
Lamb, M. E., Orbach, Y., Sternberg, K. J., Hershkowitz, I., & Horowitz, D. (2000). Accuracy of investigators’ verbatim notes of their forensic interviews with alleged child abuse victims. Law and Human Behavior, 24(6), 699–708. https://doi.org/10.1023/A:1005556404636
Lewinsohn, P. M., Flippo, J. R., & Bergquist, W. H. (1970). Leveling-sharpening: Its relation to repression-sensitization and memory. Psychological Reports, 27(2), 211–214. https://doi.org/10.2466/pr0.1970.27.2.211
Marsh, E. J. (2007). Retelling is not the same as recalling: Implications for memory. Current Directions in Psychological Science, 16(1), 16–20. https://doi.org/10.1111/j.1467-8721.2007.00467.x
Marsh, E. J., & Tversky, B. (2004). Spinning the stories of our lives. Applied Cognitive Psychology, 18(5), 491–503. https://doi.org/10.1002/acp.1001
Mesoudi, A., & Whiten, A. (2004). The hierarchical transformation of event knowledge in human cultural transmission. Journal of Cognition and Culture, 4(1), 1–24. https://doi.org/10.1163/156853704323074732
Mesoudi, A., & Whiten, A. (2008). The multiple roles of cultural transmission experiments in understanding human cultural evolution. Philosophical Transactions of the Royal Society B: Biological Sciences, 363(1509), 3489–3501. https://doi.org/10.1098/rstb.2008.0129
Ost, J., & Costall, A. (2002). Misremembering Bartlett: A study in serial reproduction. British Journal of Psychology, 93(2), 243–255. https://doi.org/10.1348/000712602162562
Ost, J., Udell, J., Dear, S., Zinken, J., Blank, H., & Costall, A. (2022). The serial reproduction of an urban myth: Revisiting Bartlett’s schema theory. Memory, 30(6), 775–783. https://doi.org/10.1080/09658211.2022.2059514
Treadway, M., & McCloskey, M. (1987). Cite unseen: Distortions of the Allport and Postman rumor study in the eyewitness testimony literature. Law and Human Behavior, 11(1), 19–25. https://doi.org/10.1007/BF01044836
Treadway, M., & McCloskey, M. (1989). Effects of racial stereotypes on eyewitness performance: Implications of the real and the rumoured Allport and Postman studies. Applied Cognitive Psychology, 3(1), 53–63. https://doi.org/10.1002/acp.2350030106
Tversky, B., & Marsh, E. J. (2000). Biased retellings of events yield biased memories. Cognitive Psychology, 40(1), 1–38. https://doi.org/10.1006/cogp.1999.0720
Wagner, U., & Echterhoff, G. (2025). Audience tuning effects on communicators’ memory: The role of the communicator’s own initial judgment. Memory, 33(7), 845–860. https://doi.org/10.1080/09658211.2025.2534147
Warren, A. R., & Woodall, C. E. (1999). The reliability of hearsay testimony: How well do interviewers recall their interviews with children? Psychology, Public Policy, and Law, 5(2), 355–371. https://doi.org/10.1037/1076-8971.5.2.355
Legal Authorities
Crawford v. Washington, 541 U.S. 36 (2004).
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999).
State v. Chapple, 135 Ariz. 281, 660 P.2d 1208 (1983).
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. Houser, 36 M.J. 392 (C.M.A. 1993).
U.S. Const. amend. VI.
Fed. R. Evid. 702.
Fed. R. Evid. 805.
Fed. R. Evid. 806.
Mil. R. Evid. 702.
Mil. R. Evid. 805.
Mil. R. Evid. 806.
Mil. R. Evid. 1102.
Ariz. R. Evid. 702.
Ariz. R. Evid. 805.
Ariz. R. Evid. 806.
Ariz. Sup. Ct. Order R-23-0004 (amending Ariz. R. Evid. 702, effective Jan. 1, 2024).
Related topic papers: Reconstruction and Recall · Schema and Script Memory · Remembering Changes Memory · External Sources of Memory Distortion · Memory for Oral Communications · Gist Memory
Related topics from Arizona Forensic Services: Child Abuse Allegations · Courts-Martial