Case File 284 · Investigation Psychology

How Can Investigators Check Whether a Confession Is Reliable?

By Murder Book Editorial TeamPublished by Gungrounds, the studio making Murder BookPublished 3 October 2026Updated 3 October 2026

A confession is not automatically reliable just because someone said it. Investigators check whether a confession is reliable by looking for independent corroboration: details the suspect provided that can be verified through physical evidence, witness accounts or forensic findings, and that were not already known to the suspect from the interrogation itself. Research by Saul Kassin and colleagues has shown that false confessions can contain accurate crime details fed by investigators, and that a confession can even corrupt other evidence in a case. The key question is not whether the confession sounds convincing, but whether its specific claims hold up when tested against facts the suspect could not have guessed.

Key takeaways

A reliable confession should contain details the suspect could only know if they committed the crime, and those details must be independently verified.
Corroboration means checking confession claims against physical evidence, forensic findings and witness accounts that exist independently of the interrogation.
Research shows false confessions often contain accurate crime details because interrogators inadvertently or deliberately share information with suspects.
Kassin, Bogart and Kerner (2012) found that confessions can corrupt other evidence, creating the appearance of corroboration where none truly exists.
Recording the entire interrogation helps reveal whether key details originated with the suspect or with the interviewer.

What does corroboration mean in the context of a confession?

Corroboration means finding evidence that supports or confirms a claim. When applied to confessions, it means checking whether the specific facts a suspect described, such as the location of a weapon, the sequence of events or the condition of a crime scene, match evidence discovered through independent investigation.

The concept matters because courts treat corroborating evidence as a safeguard. Appeals courts may uphold a conviction even when questions arise about how a confession was obtained, provided other evidence independently supports the suspect's guilt. But as Kassin noted in a 2012 study, "there may well be the appearance of corroboration, but it is false evidence that was corrupted by the confession, not independent at all."

Why can a false confession contain accurate details?

One of the most counterintuitive findings in confession research is that false confessions frequently contain details that match the crime. Drizin and Leo (2004), writing in the North Carolina Law Review, described a three-step process: misclassification (deciding the wrong person is guilty), coercion (pressuring a confession), and contamination (transferring crime-scene details to the suspect during interrogation).

Contamination happens when investigators share facts about the crime, sometimes deliberately as part of a confrontational technique, and sometimes without realising it. The suspect then repeats those facts in the confession, making it appear that they have guilty knowledge. This is why researchers distinguish between details the suspect volunteered unprompted and details that were introduced during questioning. For more on how confessions can absorb crime-scene information, see What Is Confession Contamination?.

How can confessions corrupt other evidence?

Kassin, Bogart and Kerner published "Confessions That Corrupt" in the journal Psychological Science in 2012. They analysed 241 DNA exoneration cases from the Innocence Project and found that cases involving false confessions were significantly more likely to contain multiple types of evidence error, including flawed forensic analysis, unreliable informant testimony and mistaken eyewitness identifications.

The study found that multiple evidence errors appeared in 69% of false confession cases, compared with fewer than half of eyewitness misidentification cases. In two thirds of the false confession cases, the confession came first, followed by the other errors. This sequence suggests a causal pattern: once a confession exists, other participants in the justice system, including forensic examiners and witnesses, may interpret ambiguous evidence in a way that confirms the confession.

Kassin described this as a forensic confirmation effect. When forensic analysts know a suspect has confessed, their analysis of ambiguous evidence can shift toward a guilty interpretation. This means the very evidence that is supposed to independently verify the confession may itself have been influenced by it. The concept overlaps with confirmation bias in investigations.

What should investigators look for when evaluating a confession?

Researchers have suggested several criteria for assessing confession reliability:

Unique knowledge. Did the suspect describe facts that were not publicly available and were not shared during interrogation? Examples include the location of hidden evidence, details about the victim's injuries that were not released, or actions that only the perpetrator would know.

Lead evidence. Did the confession lead investigators to new evidence they did not already have, such as a discarded weapon or stolen property?

Consistency with physical evidence. Do the suspect's described actions match forensic findings, such as wound patterns, blood distribution or DNA evidence?

Internal consistency. Is the account coherent and free of contradictions that would suggest the suspect is guessing?

Absence of contamination. Can the interrogation record show that key details originated with the suspect rather than the interviewer? This is one of the strongest arguments for recording entire interrogations.

Kassin (2017), writing in Wiley Interdisciplinary Reviews: Cognitive Science, emphasised that although police are trained to seek corroboration, the criteria they use may not adequately protect against false confessions, because most false confessions already contain accurate details and other content that appears credible.

Why does recording the interrogation matter?

A complete recording of an interrogation, from start to finish, provides the only reliable way to determine who introduced specific details. If a suspect mentions the calibre of a weapon before any investigator mentioned it, that is stronger evidence of guilty knowledge than if the interviewer raised the calibre first.

Without a recording, there is no way to reconstruct this sequence after the fact. Memory for conversation details is unreliable, and written summaries typically omit the back-and-forth exchanges where contamination occurs. Research on why false confessions can contain accurate crime details underscores this point.

Can a voluntary confession still be unreliable?

Yes. Courts currently require that a confession be voluntary to be admissible, meaning it was not obtained through physical coercion or direct threats. But a confession can be voluntary and still unreliable. A suspect may confess voluntarily for reasons unrelated to guilt: to protect someone else, to end a stressful situation, or because they have come to believe they may have committed the crime. Research on voluntary false confessions describes several documented examples.

The distinction between voluntariness and reliability is important because the legal system has historically focused on the first while paying less attention to the second. A confession that a suspect gave willingly still needs to be checked against independent evidence.

How does this connect to Murder Book?

Murder Book: The Art of Interrogation is an upcoming narrative interrogation game from Gungrounds that puts players in the position of evaluating what suspects say. Development screenshots show an evidence board where statements such as "I NEVER met her!" are linked to physical evidence, requiring players to check whether claims hold up or fall apart under scrutiny.

The game does not treat a confession or admission as the end of the investigation. Players examine unreliable stories, look for narrative gaps and contradictions, and weigh behavioural cues as reasons to dig deeper rather than as proof. This mirrors the research principle that corroboration must be genuinely independent, not shaped by the very statement it is supposed to verify.

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