How likely are you to believe you saw something that didn’t happen? Depends on the circumstances, of course.

The January 2009 issue of Psychological Science includes a study titled “Recalling a Witnessed Event Increases Eyewitness Suggestibility: The Reversed Testing Effect”. 

Studies have already shown that receiving misinformation about witnessed events prior to recall would distort perception. The abstract of the article starts with this documented fact:

People’s later memory of an event can be altered by exposure to misinformation about that event. The typical misinformation paradigm, however, does not include a recall test prior to the introduction of misinformation, contrary to what real-life eyewitnesses encounter when they report to a 911 operator or crime-scene officer.

So if you are given false information about what you saw before you start to tell your version… you’re likely to mess up the “details”. Common sense stuff. But, as is pointed out, that’s not likely to happen at a crime scene.

The authors of the article initially hypothesized that if you gave misinformation to two different sets of people about something they witnessed, letting one group tell their version before being given false details would enhance the reliability of their later recall:

Because retrieval is a powerful memory enhancer (the testing effect), recalling a witnessed event prior to receiving misinformation about it should reduce eyewitness suggestibility.

Again, it’s common sense. Who hasn’t experienced that sense of urgency? When you know you want to remember something, you start getting the details down. Maybe you tell someone; maybe you just start repeating the stuff in your head. But everyone knows that the sooner you try to recall something, the better you’ll remember it.

Except it ain’t so. The study proved just the opposite. (You’ve got to love science. The whole point of proving something using the scientific method can lead to unexpected results.)

The article is pay-per-view, but a report in Science Daily gives us some more details:

A group of volunteers watched the first episode of "24" and then either took an immediate recall test about the show or played a game. Next, all of the subjects were told false information about the episode they had seen and then took a final memory test about the show…

The researchers found that the volunteers who took the test immediately after watching the show were almost twice as likely to recall false information compared to the volunteers who played the game following the episode.

I had to read that several times for it to sink in. You watch “24”. You take a test about what you saw. Someone else, instead of immediately being asked what they saw on “24” does something completely unrelated.

Afterwards, both of you are given false information about what you watched. And then you are both tested on what you saw. You are re-tested; he’s asked for the first time.

Here’s the kicker… you are more likely than the other guy to repeat the false information:

The results of a follow-up experiment suggest that the first recall test may have improved subjects’ ability to learn the false information – that is, the first test enhanced learning of new and erroneous information.

I wonder if this is related to the concept that the more certain of something you are, the less likely you are to be right about it. It wouldn’t surprise me if it turned out that folks who had been quizzed immediately felt surer about their recall than ones who weren’t. But I wouldn’t know how to test that, nor do I know whether this study attempted to.

So back to the crime scene scenario, how does this play out. Well, as the authors point out, this is the crime scene scenario… most of the time. The first paragraph of the Science Daily article asks:

For example, if you witness a man in a blue sweater stealing something, then overhear people talking about a gray shirt, how likely are you to remember the real color of the thief’s sweater?

Well what if you hadn’t focused on the color of the sweater? You just saw a drug-deal-gone-bad-murder. You tell the investigators on the scene what you saw. A well meaning policeman, who had just seen Joe-the-Marijuana-Dealer in a gray shirt earlier, might well ask you, after you told him what happened, “Was he wearing a gray shirt?”

The cop might be surprised that Joe would do such a thing. After all, he’s just the guy in the neighborhood who sells a little weed. As far as he knows, Joe has no involvement in anything violent. But Joe’s the only guy he knows on the block, and when he saw Joe earlier, he was in a gray shirt.

And then, because, not in spite of, but  because you have already started recounting your recollection you are more likely to falsely implicate Joe? And to repeat that mistake later?

Scary stuff.