Memory illusions and false memories

People who heard the word "smashed" instead of "hit" later judged a filmed car crash as faster, and more than twice as many claimed to have seen broken glass that was never there.

Memory is not a video file played back. Each time you remember, you rebuild the event from pieces, and the rebuilding can bring in details from elsewhere. Most of the time this works well enough. Sometimes it produces memories that feel vivid and certain and are wrong. This unit covers the main lab findings and lets you test yourself with one of them before reading how it works.

Try this first

Read the following words once, at a steady pace, then cover them and answer the first three questions in the quiz at the bottom of this unit before reading on: bed, rest, awake, tired, dream, wake, snooze, blanket, doze, slumber, snore, nap, peace, yawn, drowsy.

Words that were never on the list

The list you just read was built like those in a 1995 study by Henry Roediger and Kathleen McDermott, adapting an older method by James Deese. Every word is linked to one word that is missing: sleep. In their first experiment, with 12-word lists, people recalled the missing word after 40 percent of the lists. With 15-word lists in the second experiment the rate rose to 55 percent. On a later recognition test, people said they had seen the missing word nearly as often as words that really were on the lists, often with high confidence. The words activate their shared associate so strongly that the mind later takes the associate for something it saw. Psychologists call this the DRM paradigm, after Deese, Roediger and McDermott.

Leading questions

Elizabeth Loftus and John Palmer showed in 1974 how a single word in a question can reshape memory. College students watched films of car accidents and were asked how fast the cars were going when they "smashed", "collided", "bumped", "hit" or "contacted" each other. Those who heard "smashed" gave the highest speeds. A week later, people from the "smashed" group were more than twice as likely to say they had seen broken glass, though the films showed none. Loftus called this the misinformation effect: information supplied after an event can change the memory of the event itself. Her work changed how police question witnesses. Neutral questions and blind photo lineups, where the officer running the lineup does not know who the suspect is, are now widely recommended.

Seven ways memory goes wrong

Daniel Schacter (2001) sorted everyday memory failures into seven "sins" in three groups. Forgetting: transience (fading over time), absentmindedness (not paying attention in the first place) and blocking (the tip-of-the-tongue state). Distortion: misattribution (getting the source wrong), suggestibility (taking in misleading information from others) and bias (current beliefs reshaping the past). Intrusion: persistence (unwanted memories that keep coming back). Schacter's point was that these are side effects of a system that is usually useful. Filling gaps with what is likely, for example, is what makes the DRM illusion possible, and it also lets you understand a half-heard sentence.

Do these findings replicate? The DRM effect is among the most reliable in memory research and has been reproduced many times with published word norms. Older classics have had a bumpier path. Frederic Bartlett's 1932 study, in which retold stories drifted toward the reteller's culture, had only failed replications until Bergman and Roediger reproduced its basic finding in 1999, partly because Bartlett had not recorded his exact instructions. A lesson for learners: confidence is not proof. When you study, check your answers against the source, and treat a strong feeling of knowing as something to test.