Most of us can't imagine confessing to a crime we didn't commit, but experts say it's easier than you might think to get someone to crack under pressure.
Saul Kassin, a psychology professor at Williams College, says that obtaining a false confession often depends on two factors: an authority figure insisting on guilt and then lying to suspects about false evidence connecting them to a crime.
"If you look at proven false confessions, in almost every one of the cases you find there was a presentation of false evidence either by falsely claiming to the suspect that he failed a lie detector test, or that the victim's grasp had his hair in it," Kassin said.
With Kassin's help, "Primetime" designed a hidden camera test to see if manufactured evidence would increase the confession rate among college students.
The lab was set up to look like many interrogation rooms -- small, cramped, stark walls, no windows, unpleasant and hot.
The college freshman recruited for the test believed they were being tested on how fast they could type. To increase stress, a metronome was set up nearby, ticking at a fairly rapid pace.
Ben, the instructor, told the students there was just one thing they could not do. "Whatever you do, don't hit the alt key because there's a glitch in the program and we're trying to get it fixed but for now it could really mess up the program," Ben told them.
But that wasn't true at all. Ben and his accomplice, Stacey, who posed as another student, tried to see if the unwitting subjects would confess to hitting the key, even when they hadn't.
When one student in the test, Quinn, was told he'd hit the alt key, his denial was tentative.
"Wait a minute, did you just hit the alt key?" Ben asked.
"I don't think so," Quinn said. Then appearing worried, he asked, "You lost everything or just this?"
First the instructor, an authority figure to a student, told Quinn he'd done something wrong. Then to test the second part of Kassin's theory, Stacey lied by saying she saw Quinn hit the key, thereby supplying the false evidence.
Quinn never resisted the accusation, and began apologizing profusely. He even signed a confession.
When Kassin entered the room to explain that it was all just a test, Quinn was relieved but then shocked.
"It's just really disorienting, and if two people are telling you something you did and say it like, 'No I saw you do this,' you just sort of have to agree with them after awhile," he said.
And Quinn was not the only one in the group to succumb: Seven of the eight test subjects confessed falsely to hitting the alt key.
In a similar study, students were told that they would have to pay hundreds of dollars to replace the computer, and several of them still falsely confessed, said Kassin.
Pressing a computer key is clearly not in the same category as committing murder, and the students also didn't face the kind of hostile atmosphere that suspects in criminal investigations often endure.
But Kassin said the principle for how to get a false confession holds. "Even as you raise the stakes, you're still going to get a certain percentage of people that get broken down by the process, and when the false evidence is presented, they confess," he said.
It's unclear, of course, whether the students in this test would confess to something more serious, but for some of them, the experience made them believe anything was possible.
"I think that if people are telling you that they have evidence and they saw you do it, or they know something about it that you apparently didn't know, I think you can confess to doing something like murder," Quinn said. "It would be that confusing."