Kelly Loudenberg's new series for Netflix, The Confession Tapes, presents a stimulating look into failures of the American criminal justice system, failures of forensic science, and failures of memory.

As a scientist who studies false confessions and false memories, and has written a book about the topic, I want to give you a crash course in some of the science behind the series, mostly using the case of Karen Boes to elucidate the concepts.

The case began when police responded to a call that a house was on fire, and found the burned body of a teenage girl inside the house. The girl’s mother, Karen Boes, was brought in for questioning shortly after they discovered the body.

Upon first arrival at the police station, Boes could be heard repeatedly denying being at the house when the fire happened. When prompted, she repeatedly and clearly stated, that she was not involved with the fire, saying outright “I did not start that fire.”

Problem 1: Innocence is not transparent

The police then ask Boes to take a polygraph, a lie detector, test, which she does willingly. In The Confession Tapes documentary she states that she did this because “there was no doubt in my mind whatsoever. I had nothing to hide.”

Up to this point, what Boes has done is what many people in this situation would have done. Waiving the right to silence, taking polygraphs when asked, and talking to police before a lawyer can be summoned is not uncommon. This is particularly the case for those who are innocent. But researchers have repeatedly shown that innocence can put innocents at risk.

For example, Kassin and Norwick conducted a mock-crime study where they found that participants were more than willing to waive their right to silence, and this was particularly the case for those who were innocent. This is because innocent people believe that there is no damning evidence to be uncovered, nothing to be lost by talking—“strikingly…participants who were truly innocent were significantly more likely to sign a [Miranda] waiver than those who were guilty. Naively believing in the power of their innocence to set them free.”

The problem is that innocence isn’t transparent, and we are naive if we assume it must be. Humans are pretty bad at intuitively knowing whether other people have something to hide. But, when an interrogator keeps hearing “I’m innocent” and they are convinced that the suspect is guilty, they get annoyed.

Investigators might even put on more pressure than if the suspect were actually guilty. In a different study. Kassin, Goldstein, and Savitsky found that when an interviewer presumes that a suspect is guilty, even though they are innocent, they put on more pressure in the interview. This can make an innocent suspect react defensively, and make them look more guilty than actual guilty people.

The researchers argue that these findings also mean that “innocent suspects increase the risk that they undergo coercive interrogations that set into motion various confirmation biases.” The interrogator thus falls prey to tunnel vision—ignoring exonerating evidence while only focusing on information that confirms their initial belief. Police in such a situation may ignore the pleas that the person is innocent, and cling to any information that supports their initial suspicion.

Lesson 1: Use your right to silence, and do not talk to the police until you have a lawyer present.

Problem 2: Bad Forensic Evidence

In the case of Boes, she “failed” the polygraph test when it appeared to show she was lying about her knowledge of the fire. Despite the fact that the polygraph is not considered by scientists to be a foolproof way to detect deception, the interviewer got in in his head that Boes was guilty. And this belief seemed unshakable. As the interviewer himself stated during the interrogation “The machine really doesn’t lie.” If the machine wasn’t lying, his logic went, it meant that she was. The problem is, the machine does lie.

Nevertheless, Boes comes to accept the results of the polygraph, and begins to question her own (the show leads us to presume) innocence.

Karen Boes: “I don’t remember. I don’t know.”

Karen Boes: “I apparently did it”

Interviewer: “What makes you say that?”Karen Boes: “The test.”

When we stop trusting our own memories of a criminal event, we open our memories up to distortion and suggestion. Researcher Gudjonsson coined the term memory distrust syndrome to describe this, which he explains as “a condition where people develop profound distrust of their memory and become susceptible to relying on external cues and suggestions from others.” It is also the main topic of my own book, The Memory Illusion. In this state we begin to accept alternate explanations. We may even come to believe, temporarily or long-term, that we committed a crime although we are innocent.

Lesson 2: There is no such thing as a foolproof lie detecting machine.

Problem 3: Leading and Suggestive Interviewing

In addition to accepting the polygraph evidence, Boes is bombarded with leading and suggestive interview techniques that are known to facilitate the creation of false memories.

For one, Boes was told that she had simply not accepted her crime, and her mind was hiding this memory from her. When denying any involvement with the death of her daughter, the interviewer responded with; “Ok. Either that or you don’t want to believe that right now... There are some things that we do that our bad mind causes us to do that our good minds would never accept.” This is not dissimilar from the contentious notion of repression, a fertile foundation for false memories.

Second, on various occasions throughout hours of interrogation, it was suggested to her that she must have been there; “You were there,” and that she must know more; “I know you know more about the fire. I know you were there. I just want you to think.” Such misinformation and suggestion can give a doubting individual something to build an imagined event on. We might begin to wonder, if I must have been there, what must I have done? Details gained from this process can firm up over time, shifting from guessing to making incorrect statements of fact.

Two other factors that make us even more prone to false memories and false confessions are (a) a long and persistent interrogation where someone is presumed to be guilty and (b) high emotional intensity—both of which are seen in this case.

Lesson 3: Write down your memory of events before you go into an interrogation, so you can later see whether your memory has changed.

This potent mix was the beginning of the end for Boes. Combined with problematic arson evidence admitted by an expert, a type of forensic science fraught with difficulty, she goes on to offer incriminating evidence which is later used to convict her. A picture is painted of a tragic story of the exhaustion of dealing with a teenaged daughter.

But is it true? How do we know that Boes didn’t do it? Certainly the interrogation tapes point to the use of many problematic police interview techniques, which call the reliability of the confession evidence generated in this case strongly into question.

Cases like this teach us that police need to know more about the fragility of memory, and the huge impact that bad forensic evidence can play. They also teach us to be more cautious, and perhaps less compliant, with police interview techniques.

There are three cases in The Confession Tapes series that include most of the problems I have discussed here. I have put together a mini-summary of each:

Case 1: Karen Boes

The sentence: Karen Boes was convicted of first-degree murder, for allegedly burning her daughter to death in the family home while her daughter slept.

The bad science: Misuse of the polygraph to establish guilt. Questionable analysis of burn marks. Use of leading and suggestive interview techniques

Lead up to the confession:

Karen Boes: “I apparently did it”

Interviewer: “What makes you say that?”

Karen Boes: “The [polygraph] test.”

Case 2: Buddy Woodall

The sentence: Buddy Woodall was convicted of murder, for his alleged involvement in the shooting of his uncle and a family friend.

The bad science: Unscientific computer voice stress test to establish guilt. Use of leading and suggestive interview techniques.

Lead up to the confession:

Buddy Woodall: “You got it set in your mind that I’m a murderer”

Buddy Woodall: “You’re trying to say stuff that I did not say”

Case 3: Wes Myers

The sentence:

Wes Myers was found guilty of first-degree murder and arson, for allegedly strangling his girlfriend to death at the bar she worked at, and then setting the room on fire.

The bad science: Fabricated evidence, that hair sample was a DNA match. Use of leading and suggestive interview techniques.

Lead up to the confession:

Wes Myers: “I was questioning myself”

Wes Myers: “They told me that I’d blacked out