Yes. People can sincerely describe the same event differently because they noticed different things, recalled different details or later encountered different information. But sincerity and accuracy are separate questions. A vivid memory can contain errors, and a cautious witness can remember something correctly.
Research helps explain why some variation is unsurprising. It cannot tell us, centuries later, what happened inside a particular Gospel writer’s memory. The Gospels also involve source use, oral transmission, literary choices and theology. We should use psychology to sharpen our questions, not announce that science has proved or disproved the Gospel accounts.
Go a little deeper
Start with an ordinary experience
Two people leave the same family gathering. One remembers an affectionate speech; the other remembers a tense exchange near the door. Both may be right about what they noticed. If they disagree about who was present, however, “different perspectives” alone may not settle the issue.
This everyday illustration helps separate three stages: what a person notices, what they later remember and what they choose to report. A written narrative adds another stage: how an author arranges material. Those stages can preserve information and introduce differences. They should not be collapsed into one explanation.
The following studies concern different tasks and conditions. Their findings should be read with those limits attached. None studied the Gospel writers.
Six studies, with the limits left in
Being there does not mean noticing everything
Simons and Chabris, 1999. Participants watched basketball passes while performing a counting task. Many did not notice an unexpected person crossing the scene. What people attended to affected what they noticed.
What this supports: different observers can register different aspects of the same scene.
Limit: a short video with an imposed task is not equivalent to following Jesus or witnessing an execution. The study cannot explain any particular Gospel omission. Original study (opens in a new tab) ↗
The question can influence the answer
Loftus and Palmer, 1974. After viewing collision films, participants gave different speed estimates depending on the verb used in a question. In a second experiment, stronger wording increased later reports of broken glass that had not appeared.
What this supports: language encountered after an event can influence a report.
Limit: altered answers do not all prove that an original memory was erased. The authors discuss response bias too. Brief film experiments do not show that witnesses usually invent entire events. Original study (opens in a new tab) ↗
A clear memory is not a perfect recording
Talarico and Rubin, 2003. Fifty-four students described hearing about the September 11 attacks and an ordinary event, then were retested after different delays. Consistency declined for both kinds of memory, while confidence and vividness remained comparatively strong for the September 11 recollections.
What this supports: feeling certain is not enough to establish every remembered detail.
Limit: these were memories of hearing the news, not witnessing the attacks. The comparison was with initial reports, not a complete independent record of what happened. Original study (opens in a new tab) ↗
People can borrow details without intending to deceive
Gabbert, Memon and Allan, 2003. Participants saw different perspectives on a staged event. After discussing it, some reported details that had been visible only to their partner.
What this supports: later discussion can influence an individual account. Agreement may reflect shared information as well as independent observation.
Limit: a detail borrowed from someone else may be accurate. This study is partly about the source of a recollection; it does not mean agreement always signals error or deception. Study and abstract (opens in a new tab) ↗
Memory can also remain substantially accurate
Yuille and Cutshall, 1986. Twenty-one witnesses gave police accounts after a shooting in British Columbia. Thirteen took part in research interviews four to five months later. The abstract reports considerable accuracy, relatively little change and limited effects from the misleading questions used, alongside errors involving colour and estimates of age, height and weight.
What this supports: consequential real events can be remembered well despite some errors.
Limit: one incident, a small voluntary follow-up group and months rather than decades cannot establish a universal rule. It does not show that stress always improves memory. Study record (opens in a new tab) ↗
Confidence can matter under known conditions
Wixted and colleagues, 2016. Their analysis of Houston police photo lineups found that initial confidence was related to estimated identification accuracy.
What this supports: it is too sweeping to say confidence tells us nothing.
Limit: accuracy required statistical estimation. The conclusion concerned adults’ first identifications in fair, double-blind lineups, where the administrator does not know the suspect’s position. It does not automatically apply to later confidence, repeated questioning or an author writing decades after an event. Original study (opens in a new tab) ↗
Apply the findings without making them do too much
These studies help challenge two tempting rules: “Honest witnesses must agree in every detail” and “Confident witnesses must be right.” They also challenge an equally sweeping scepticism: “Memory changes, so remembered events have no evidential value.” What matters is the task, the conditions and the particular claim.
Our inference for Gospel study is therefore limited. Some differences are compatible with honest testimony. Other differences could reflect error, shared tradition or authorial choices. The laboratory findings do not identify which explanation applies to a particular verse.
The time intervals matter. A report collected soon after an event differs from a narrative written decades later that may include earlier written or oral sources. Repetition can preserve a story, but repeated discussion can also alter it. We need evidence about transmission rather than assuming it was either perfectly controlled or hopelessly corrupt.
There is also a difference between memory and the interpretation of an experience. Someone may accurately report having an experience while other people disagree about its cause. Assessing claims of appearances after Jesus’ death involves testimony, interpretation and competing explanations. These studies did not test resurrection claims and cannot adjudicate them.
A common-sense way to weigh an account
- Identify the claim. “An execution happened” is different from “these were the exact last words.”
- Ask how the information reached us. Was the writer present, using an earlier source or reporting a shared tradition? What evidence supports that answer?
- Look for independent checks. Several accounts copied from one source do not create several independent confirmations.
- Separate omission from incompatibility. Not mentioning a person differs from saying that person was absent.
- Test the proposed explanation. Is a reconciliation supported by the wording, or is it merely possible?
- Match confidence to the evidence. A broad historical conclusion can be strong while some details remain uncertain.
We can take human testimony seriously without treating human memory as infallible.
Read the comparison of the crucifixion accounts to see how these questions work with actual Gospel passages.
Sources & further reading
Scripture links open the passage or its chapter. Research links identify the underlying studies; historical and theological interpretations are distinguished in the article. The summaries are editorial explanations, not quotations from the researchers.
- Simons, D. J., and Chabris, C. F. (1999). Gorillas in Our Midst. Perception 28, 1059–1074. DOI: 10.1068/p281059. Original paper. (opens in a new tab)
- Loftus, E. F., and Palmer, J. C. (1974). Reconstruction of Automobile Destruction. Journal of Verbal Learning and Verbal Behavior 13, 585–589. DOI: 10.1016/S0022-5371(74)80011-3. Original paper. (opens in a new tab)
- Talarico, J. M., and Rubin, D. C. (2003). Confidence, Not Consistency, Characterizes Flashbulb Memories. Psychological Science 14, 455–461. DOI: 10.1111/1467-9280.02453. Original paper. (opens in a new tab)
- Gabbert, F., Memon, A., and Allan, K. (2003). Memory Conformity. Applied Cognitive Psychology 17, 533–543. DOI: 10.1002/acp.885. Publisher abstract and article. (opens in a new tab)
- Yuille, J. C., and Cutshall, J. L. (1986). A Case Study of Eyewitness Memory of a Crime. Journal of Applied Psychology 71, 291–301. DOI: 10.1037/0021-9010.71.2.291. Bibliographic record; our summary uses the original abstract and opening pages. (opens in a new tab)
- Wixted, J. T., Mickes, L., Dunn, J. C., Clark, S. E., and Wells, W. (2016). Estimating the Reliability of Eyewitness Identifications from Police Lineups. PNAS 113, 304–309. DOI: 10.1073/pnas.1516814112. Original paper. (opens in a new tab)
Research and source checks: October 2026. These articles develop the revised questions supplied for The Jesus Inquiry. No specialist peer review is claimed.