Cognitive biases
The misinformation effect: why we “remember” what we were told
One morning Irina walks into a new bookstore on the corner and buys a book. In the evening a friend asks: “Is that the shop with the little table by the entrance with a red tablecloth? Is it cozy there?” Irina answers: “Yes, I think it’s cozy.” A week later she recommends the shop to a coworker: “There’s such a cute little table with a red tablecloth by the entrance.” The coworker goes in and finds no table. The clerk is surprised: they have never had a table by the entrance. Irina was not lying. She truly felt she had seen that table.
This is the misinformation effect: if after an event we hear an incorrect detail about it, we may later easily recall it as part of what we saw. And tell it with confidence.
Below, we look at what research has shown, why this happens, and how to remember more accurately. In the middle of the article you will take Irina’s place: look at a morning outside a bookstore and test your memory.
Contents
In short: how the misinformation effect works
What the misinformation effect is
The best-known experiment was done by the American psychologists Elizabeth Loftus and John Palmer (a 1974 paper). Students watched short films about car accidents and then answered questions. The key question was worded differently for different people. Some were asked how fast the cars were going when they “smashed into” each other. Others, when they “hit.” Those who heard “smashed into” on average gave a higher speed.
A week later the participants were asked: “Did you see any broken glass?” There was no broken glass in the film. Among those who heard “smashed into,” about one in three said yes. Among those who heard “hit,” about one in seven. One word in a question changed not only the answer to it but also what people “remembered” a week later.
Such experiments have since been repeated many times. In a 1978 paper, people watched slides of an accident and then received information about it, including incorrect information. The incorrect detail made it harder afterward to recognize what had actually been there. And it interfered more if it was given not right after the event but shortly before the test.
Suggested details are not recalled timidly. In another experiment, participants chose the suggested detail as quickly and confidently as something they had really seen. In a 2005 review, Loftus sums up thirty years of such research: the effect has been found many times, and some people give in to it more than others.
Try it yourself: a morning outside the bookstore
Now you are in Irina’s place. The game takes a couple of minutes. Here is what will happen.
Step 1. For ten seconds you look at a picture: a morning, a street corner by a bookstore. Just look, the way you look on your way to a shop.
Step 2. In the evening a friend asks Irina about the shop. Four short questions, one per screen. Answer as you remember.
Step 3. In a list, you mark everything you think was in the picture.
The questions differ a little from player to player: the computer picks at random which set you get. Where the difference comes from will become clear at the end. Why is the picture shown for such a short time? In life, too, we rarely study a street on purpose. We usually just walk past, and the questions come later.
What usually happens. In one of Loftus’s experiments, people watched a video of a white sports car driving along a country road. Half were asked: “How fast was the white sports car going when it passed the barn?” There was no barn in the video. The others were asked the same question without the barn. A week later everyone was asked: “Did you see a barn?” About 17% of those who had the barn mentioned said yes, and about 3% of the others. In the same paper, a detail mentioned in passing worked more strongly than a direct question like “Did you see…?”
What your result can and can’t show. Each player tests only one suggestion, and it can be marked by guessing. If you marked the suggested item and not the other one, that looks like what the scientists saw. If the suggestion did not take, that is a common result: in the barn experiment, most people did not “remember” the barn either. Besides, you read this article and most likely expected a trick. And minutes passed between the picture and the test, not a week.
When it is not a mistake. Relying on other people’s accounts is normal: often they help us remember more accurately. The problem is something else. Later we can’t tell the suggested from the seen, and we tell it just as confidently.
Why a suggestion passes itself off as a memory
Imagine you are putting together a jigsaw puzzle with a few pieces missing. Someone hands you a similar piece, and it fits. A week later it is hard to say which piece was in the box from the start. This is an image, not the exact workings of memory. What actually happens, scientists are still arguing about.
The suggestion blends with what was seen. Loftus and her colleagues thought this: what we hear after an event merges with what the person saw. Later they can no longer separate which detail came from where.
The original memory is intact, it is just missing. The psychologists McCloskey and Zaragoza objected: perhaps the suggestion spoils nothing. In their view, if a person did not remember a detail, they may simply choose the one they heard. In their experiments, which were set up differently, the suggestion did not affect memory for the event itself.
We mix up where we know things from. Other researchers asked participants not only what they remembered but also where from. People sometimes thought they remembered seeing something they had only been told about. But not always: how strongly this showed depended on the conditions.
However it works on the inside, from the outside the result is the same: a person tells what was suggested as if they had seen it.
Where it gets in the way
When people are questioned about what happened. The question “Did you see him run the red light?” already contains the answer. It is no accident that in court there is such a thing as a “leading question.” Witnesses to an accident often discuss it while waiting for help, and afterward their accounts can easily become alike. This is an everyday observation.
Relationships. After an argument a friend sympathizes: “He yelled at you, didn’t he?” A month later you are sure he yelled, though he may not have raised his voice. This is an everyday observation, not the result of an experiment. But it works the same way as the question about the barn.
Work. A coworker retells a meeting: “We did agree to submit it on Friday.” A week later the whole team remembers “Friday,” though no deadline was agreed at the meeting.
News and social media. You saw a video of an incident, then read the caption and the comments: who was wearing what, who shouted what. A few days later you retell the video together with those details, though they were not in the video itself.
How to remember more accurately
The misinformation effect is one of the biases on the cognitive bias map, in the section on memory. Right there is the illusory truth effect: what we have heard many times seems truer.
Frequently asked questions
If we know about the effect, does it stop working?
Not quite. In her review, Loftus writes that warnings sometimes help, but only under certain conditions. In a summary of 25 studies, people were warned after the suggestion that what they had heard might contain errors. On average such a warning weakened the effect by more than half. Whether knowing about the effect in advance protects us, those studies did not test.
How is this different from hindsight bias?
With hindsight bias, once we know the outcome, we remember our earlier opinion differently: “I knew it all along.” With the misinformation effect, the account of the event itself changes: details appear in it that other people suggested to us.
Does that mean we shouldn’t rely on other people’s accounts?
No. Relying on other people’s accounts is normal: often they help us remember more accurately. The problem is something else. Later we can’t tell the suggested from the seen, and we tell it just as confidently.
