Finance & investing
Investors may recall a bullish forecast about a stock but misattribute it to a credible analyst when it actually came from an anonymous forum post, leading them to place unwarranted confidence in the prediction.
Correctly remembering information but incorrectly remembering where, when, or how it was originally learned.
Imagine you hear a really funny joke, but a few days later you can't remember if your best friend told it to you or if you saw it on YouTube. You remember the joke perfectly, but who told it got all mixed up in your head. Your brain kept the joke but lost the label saying where it came from.
Source confusion occurs when the brain successfully stores the content of an experience — a fact, a face, a story — but fails to preserve the contextual tag linking that content to its origin. People may remember hearing a piece of information but confuse whether they read it in a newspaper, saw it on TV, or heard it from a friend. This error is especially common when multiple sources share perceptual or contextual similarities, when significant time has elapsed, or when cognitive load at encoding was high. The bias reveals that human memory is a reconstructive process rather than a recording device: the brain infers the source of a memory based on qualitative features like vividness, emotional tone, and perceptual detail, and these inferences are frequently wrong.
The same glitch looks different depending on the terrain. Finance, medicine, a relationship, a team — same mechanism, different costume.
Investors may recall a bullish forecast about a stock but misattribute it to a credible analyst when it actually came from an anonymous forum post, leading them to place unwarranted confidence in the prediction.
Patients frequently report symptoms or medication effects they read about online as things their doctor told them, complicating clinical history-taking and potentially leading to incorrect treatment decisions.
Students may reproduce ideas from a lecture or textbook in essays without citation, genuinely believing the ideas are their own original thoughts — a form of unintentional plagiarism rooted in source confusion rather than dishonesty.
People may misremember which partner or friend made a particular promise or said something hurtful, causing conflict when they confront the wrong person about a statement that individual never made.
Users exposed to multiple onboarding tutorials, help articles, and tooltips often misattribute instructions from one platform to another, leading to frustration when expected features do not exist in the product they are currently using.
Managers may attribute a good idea to the wrong team member during performance reviews, crediting the person who presented it rather than the person who originally proposed it in an earlier meeting.
Voters may remember a claim from a satirical news source or partisan blog as coming from a mainstream outlet, lending false credibility to misinformation and shaping their political judgments accordingly.
Marcia K. Johnson and Carol L. Raye introduced the concept of 'reality monitoring' in 1981, distinguishing internal from external memory sources. Johnson, Hashtroudi, and Lindsay formalized the broader 'Source Monitoring Framework' in a landmark 1993 Psychological Bulletin paper.
In ancestral environments, remembering the content of survival-relevant information (e.g., which berries are poisonous, where predators lurk) was far more critical than remembering the exact source of that knowledge. Brains that prioritized gist-level semantic storage over precise contextual tagging could process and act on information more rapidly, even if the cost was occasional misattribution of the source.
Large language models routinely exhibit source confusion by generating text that blends and misattributes information across training documents. An LLM may attribute a quote to the wrong author, conflate findings from separate studies, or present fictional citations as real — a form of 'hallucination' that mirrors human source monitoring failures. Training data provenance is lost during the learning process, making accurate source attribution structurally difficult for these models.
50 cards are free to preview. Buyers unlock the rest of the deck plus the interactive training — Spot-the-Bias Quiz unlimited, Swipe Deck with spaced repetition, My Blindspots, Decision Pre-Flight, the Printable Deck + Cheat Sheets, and the Field Guide e-book. $29.50$59.