What does it mean to have a gut feeling that you remember something? You see someone you recognize in a coffee shop. Do you remember her from high school? Or maybe you saw her on television. Could she be the manager of your local bank? Perhaps you don’t know her at all … but you’ve still got a feeling you do. What’s that all about?
One theory of memory proposes that what we remember depends on our expectations. We’re less likely to remember our old classmate at the coffee shop than at the high school reunion. At the bank, we might greet the manager by name, but we only get a vague sense of recollection when we see her in the checkout line at the grocery store. So what cues that sense of expectation? If the grocery store pipes in the same music they play at the bank, will we remember her then? What if it turns out she’s not the bank manager, but another woman about the same height who happens to own the same blazer? Does the music help us notice the difference, or just make us more likely to ask a complete stranger about the status of our mortgage application?
Stephen Goldinger and Whitney Hansen designed a nifty experiment to see how a subtle cue affects memory. They rigged a chair with a set of speakers under the seat: they wanted to generate a subliminal signal—a “buzz” soft enough that most people claim they can’t feel it, but loud enough that people can guess that a sound might have been played. In a preliminary test, they gradually decreased the intensity of the buzz until 75 percent of participants claimed no awareness, but 65 percent still guessed correctly.
Next, they asked a new set of volunteers to memorize 96 words—48 “easy” words that had a clear visual image associated with them (e.g. “snowball”), and 48 “hard” words for abstract concepts (e.g. “movement”). After a short period where they were distracted with an opinion questionnaire, they were tested on their memory. For the test, a word would be presented on-screen, and the respondent indicated whether it was old or new—whether it had appeared in the original list of words or not. Half the time during the testing phase, the subliminal buzz was played, and half the time it was not. The same test was repeated with photos of objects, and again with faces.
For easy objects, Goldinger and Hansen found no difference between cases when the buzz was played and when it was not. This makes some sense—to take an obvious example, I’d recognize my wife and kids regardless of the context. However, for hard objects, the result was quite different:
The chart on the left compares accurate “hits”—when participants correctly identified “old” items—and “false alarms,” when new items were incorrectly identified as old. When the buzz was played, both hits and false alarms went up for hard items. Rather than actually enhancing memory, the buzz just made people more likely to say they’d seen an item before. The participants were also asked to rate their confidence in their answer on a scale of 1 to 7. The buzz increased confidence for false alarms, but decreased confidence for hits.
As a control, Goldinger and Hansen also ran the same experiment, but changed the buzz so that it was clearly audible, telling participants that they’d hear a buzz sometimes but offering no other explanation (this must have led to some interesting conversation among student participants on campus after the experiment was completed). In this case, the buzz had no impact on accuracy, either for hits or false alarms.
Goldinger and Hansen explain the data this way. When we’re certain about a memory, we aren’t influenced by external cues—subtle or otherwise. But when a memory is less sure, we tend to rely on “gut feelings.” Sometimes, however, a gut feeling is nothing more than a barely perceptible stimulus. Our gut can mislead us—and it can also be imitated by a speaker hidden inside our chair, or the caffeine buzz from an extra cup of coffee, or any number of other things.
Goldinger, S.D., Hansen, W.A. (2005). Remembering by the Seat of Your Pants. Psychological Science, 16(7), 525-529. DOI: 10.1111/j.0956-7976.2005.01569.x