Case File 069 · Investigation and Psychology

Why Innocent People Sometimes Confess

What research by Saul Kassin and others shows about why innocent people confess, who is most at risk, and why false confessions are so persuasive.

By Murder Book Editorial TeamPublished 16 September 2026Updated 16 September 2026

Innocent people sometimes confess because the situation, not the truth, becomes the deciding factor. Research led by psychologist Saul Kassin and colleagues identifies three broad routes. Some people confess voluntarily, without police pressure. Some comply: they know they are innocent but confess to escape a stressful interrogation or because they expect a better outcome. Others internalise the accusation and come to believe they may have done it. The risk is higher for young people, people with intellectual disabilities or mental illness, and people who are highly compliant or suggestible. It also rises with certain interrogation tactics, such as very long sessions, presenting false evidence and minimising the seriousness of the crime. The problem is not rare in wrongful conviction cases: the Innocence Project reports that 29% of the DNA exonerations in its data involved a false confession.

Key takeaways

False confessions are usually grouped into three types: voluntary, compliant and internalised.
Risk factors include youth, intellectual disability, mental illness, and tactics such as long interrogations, false evidence and minimisation.
In research, innocent people were more willing than guilty people to waive their rights and talk, because they believed their innocence would be obvious.
Confessions are highly persuasive; mock jurors were influenced by confessions even when they judged them to be coerced.
The Innocence Project reports that 29% of the DNA exonerations in its data involved false confessions.

How common are false confessions?

Nobody knows the true rate, because most false confessions are never discovered. What researchers can study are cases where innocence was later established.

The Innocence Project's page on DNA exonerations in the United States reports that 29% of the 375 DNA exonerations in its data involved false confessions. Of those who falsely confessed, it reports that 49% were 21 or younger at the time of arrest, 31% were 18 or younger, and 9% had mental health or mental capacity issues known at trial. The page draws on data collected up to 2020, so the figures reflect that snapshot.

In a 2008 review in Current Directions in Psychological Science, Kassin cited earlier analyses finding that 20 to 25% of DNA exonerees had confessed, and noted that known cases probably represent only a fraction of the total.

What are the three types of false confession?

In 1985, Saul Kassin and Lawrence Wrightsman proposed a three-part typology that is still widely used, although other researchers have since refined and extended it. Kassin summarises the types as follows.

Type	What happens	Why it happens
Voluntary	A person claims responsibility without prompting from police	Reasons include a desire for attention, guilt about something else, delusions, or protecting someone
Compliant	A suspect confesses while privately knowing they are innocent	To escape a stressful situation, avoid a threatened outcome or gain an implied reward
Internalised	A vulnerable suspect comes to believe they committed the crime	Exposure to highly suggestive interrogation, often including claims of evidence

Compliant false confessions are described as an act of public compliance: in the moment, the short-term relief of confessing seems to outweigh the long-term costs.

Why are innocent people interrogated in the first place?

Kassin's 2008 review describes a sequence that starts before any confession. An interrogation is often preceded by an interview in which investigators decide whether they believe the suspect. That judgement is less reliable than it feels.

A meta-analysis by Charles Bond and Bella DePaulo found that people judge truth and deception correctly only about 54% of the time on average. Kassin's review describes a study in which participants trained in a popular lie-detection method were less accurate, more confident and more inclined to see deception than untrained participants. So an innocent person who seems nervous or evasive can be misjudged, which is why behaviour on its own is not proof. The article Can You Actually Tell When Someone Is Lying? covers this research in more depth.

Once investigators believe a suspect is guilty, that belief can shape the rest of the process, a pattern explored in how confirmation bias can affect an investigation.

What situational factors raise the risk?

The 2010 white paper Police-Induced Confessions: Risk Factors and Recommendations, by Kassin, Drizin, Grisso, Gudjonsson, Leo and Redlich and published in Law and Human Behavior, identifies interrogation tactics that can increase the risk of false confessions. Among them are:

Excessive interrogation time. Very long sessions are listed as a risk factor in their own right.
Presenting false evidence. Claiming that evidence of guilt exists when it does not. Kassin's 2008 review notes that American police are permitted to do this.
Minimisation. Offering sympathy or moral justification, which can imply leniency without an explicit promise.

Laboratory research described in Kassin's 2008 review shows why these matter. In one experiment, students were falsely accused of causing a computer to crash by pressing a forbidden key. All were innocent. When a confederate claimed to have seen them press it, the share who signed a confession rose from 48% to 94%. In another study, both minimisation remarks and an explicit offer of leniency increased true confessions from guilty participants but also false confessions from innocent ones.

These are experiments with low stakes, and their results cannot be carried directly into real cases. But they show the mechanisms researchers are concerned about. Rules on what police may do differ between countries; the research summarised here is mostly from the United States.

Who is most vulnerable?

The white paper names suspect characteristics that increase risk, including adolescence, intellectual disability, mental illness and certain personality traits. Kassin's 2008 review expands on these:

Compliance. People eager to please or to avoid confrontation may go along with an accusation.
Suggestibility. People whose memories are easily altered by misleading questions or negative feedback are more open to influence. This links to the research on open-ended versus leading questions.
Psychological distress. High anxiety, fear, depression or other disorders can increase the risk.
Youth. Adolescents tend to focus on immediate relief and underweight future risks, so confessing can seem like a quick way out.

The Innocence Project's figures on the age of false confessors point in the same direction.

Why might innocence itself be a risk?

Kassin calls this the phenomenology of innocence. People who know they are innocent often believe the truth will be obvious to everyone else. As a result, they may cooperate fully and speak freely, sometimes without realising they are under suspicion.

In a study by Kassin and Norwick described in the 2008 review, participants who were innocent of a mock theft were more likely to waive their rights and talk than those who were guilty: 81% compared with 36%. Many explained that they had nothing to hide. The finding is not a recommendation about what anyone should do; it shows how a sincere belief in one's own innocence can lower a person's guard.

This article is educational and is not legal advice.

Why are false confessions so persuasive?

Confessions carry great weight with judges and juries. Kassin's review describes a study in which laypeople and police investigators watched true and false confessions recorded by prison inmates. Neither group did well, with accuracy between 42% and 64%.

In a mock-jury study, a high-pressure confession raised the conviction rate from 19% to 47%, even though participants rated the confession as involuntary and said it had not influenced them.

A confession can also colour how other evidence is seen. In a 2014 study, Jeff Kukucka and Saul Kassin found that participants who were told a defendant had confessed were more likely to wrongly conclude that two handwriting samples came from the same person.

The 2010 white paper ends with a strong recommendation for mandatory electronic recording of interrogations, so that courts can see how a confession was obtained. It also discusses other possible reforms and ways to protect vulnerable suspects. Kassin's 2008 review adds the role of expert testimony in helping courts understand confession evidence.

A related question is how a fact-finding interview differs from an accusatory interrogation, covered in interview versus interrogation.

How does this connect to Murder Book?

Murder Book: The Art of Interrogation is an upcoming narrative interrogation game from Gungrounds in Zagreb, inspired by real police footage and psychological interviews. Development screenshots show a technique menu that includes Build Trust, Bluffing, Baiting and Confrontation, and a case select screen whose results include "not guilty".

The research on false confessions is a useful counterweight for anyone playing an interrogation game. A suspect's stress, hesitation or eventual admission does not settle what really happened, and pressure can produce answers that feel conclusive but are not. Murder Book's cases are fiction, and its techniques are game mechanics, not a model of good practice. Reading its characters well means keeping uncertainty in view: behaviour is a clue to weigh, never proof on its own. Learn more in What Is Murder Book?

Murder Book is in development for PC (Steam) and mobile, with release currently planned for the end of Q2 2027.