Skip to content
Search resources Open menu

False Confessions

Secondary Confessions: Informants, Co-Accused, and Jailhouse Statements

Published

Snap Shot

  • The question: How reliable is testimony that the accused confessed to someone else, such as a cellmate, acquaintance, or co-accused, and how do factfinders and the law treat it?
  • What the research shows: In mock-jury studies, testimony that the defendant confessed to an informant or accomplice increased guilty verdicts, and in some mock-jury studies disclosure of the witness’s incentive had little effect on verdicts, although one study found that it did. In laboratory studies with students and in a hypothetical-choice study with community members, incentives increased willingness to report a confession that had not occurred. Informant testimony appears among the contributing factors in documented wrongful convictions, and false confessors have sometimes implicated other people.
  • What it does not show: No study measures how often informant or co-accused testimony is accurate in real cases. Some informants and cooperating witnesses provide accurate information that leads to evidence, and the laboratory studies use students, online participants, and hypothetical choices.
  • The law: Once the right to counsel has attached, the government may not deliberately elicit statements through an informant (Massiah, Henry), although a passive listener may report what he hears (Kuhlmann), and Miranda does not apply when the suspect does not know he is speaking to an agent (Perkins). Promises to a witness must be disclosed (Giglio); a non-testifying co-defendant’s confession naming the defendant raises confrontation problems (Bruton); in courts-martial, Article 31(b) and Mil. R. Evid. 305 reach knowing agents, and Mil. R. Evid. 306 governs joint trials.
  • Issues for each side: For prosecutors, the research favors corroborating the informant’s account with evidence the informant could not have obtained elsewhere and by full disclosure of benefits; for defense counsel, tracing what the informant could have learned from other sources and how the account developed.

A secondary confession is testimony by another person that the accused admitted the crime to him or her. The witness may be a jail or barracks cellmate, a friend or family member, a person who later became a cooperating witness, or a co-accused who describes the accused’s role while explaining his own. Researchers adopted the term to distinguish these reports from a primary confession, which the accused makes directly to investigators and which is often recorded (Neuschatz et al., 2008). A secondary confession is usually reported from memory, after the fact, by a witness who may have something to gain.

These statements reach court because they can supply the admission that a case otherwise lacks. Investigators and prosecutors depend on cooperating witnesses in many serious cases, and some provide information that is accurate and leads to evidence. At the same time, testimony from informants with incentives is among the factors associated with documented wrongful convictions, and the factfinder often has no recording of what the accused actually said.

This paper summarizes the research on how secondary confessions are produced and how jurors evaluate them, and on co-accused statements, and it describes the federal, military, and Arizona law that governs the use of informants and of one accused’s statements at another’s trial. Corroboration of confessions generally is addressed in “Corroboration and Assessing the Reliability of a Confession.”

Informants in Documented Wrongful Convictions

Informant testimony appears repeatedly in analyses of wrongful convictions. A 2004 study by Northwestern University’s Center on Wrongful Convictions reported that informant testimony contributed to 45.9% of the 111 death-row exonerations it counted (Warden, 2004). In an analysis of DNA exoneration case files, false-confession cases were more likely than eyewitness cases to contain additional errors; in order of frequency these were invalid or improper forensic science, mistaken eyewitness identifications, and informants, and where several errors were present the confession was usually obtained first (Kassin et al., 2012). The authors read this pattern as evidence that a confession can influence the evidence that follows it, including what informants report, and the 2025 scientific review lists informants among the sources of evidence that a confession can taint (Kassin et al., 2025).

Like all exoneration data, these figures come from cases known to have ended in wrongful conviction, chosen because they could be proven, and they describe how errors occurred in those cases. They do not indicate what proportion of informant testimony in ordinary cases is accurate. No study has drawn a representative sample of informant testimony and determined which accounts were true.

How Secondary Confessions Are Produced

Laboratory research has examined the conditions under which people will report a confession that was never made. In two studies with 332 college students using a computer-crash paradigm, participants were asked to sign either their own confession or a statement that another participant had caused the crash. Some participants signed the secondary confession, and the combination of false evidence and an incentive increased the rate of signing (the percentages were not available in the sources consulted for this paper); in the second study, an incentive increased secondary confessions only when the other person had denied the act, meaning the incentive increased false reports specifically (Swanner et al., 2010). The authors drew implications for the use of incentives in interviewing informants. The task was low-stakes, and signing a statement in a laboratory differs from testifying in a criminal trial.

More recent studies have asked community members to imagine themselves as inmates offered benefits for testifying. In two studies (N = 99 and N = 289), about 27% and 17% of participants were willing to testify falsely against another inmate in return for offers such as sentence reductions, immunity, or financial support, and those who were willing described their decisions mainly in terms of self-interest (Jenkins et al., 2023). Hypothetical decisions may not predict what people do when actual benefits are at stake, in either direction.

A secondary confession can also shape how other evidence is judged. Laboratory studies have found that knowledge of a confession can influence eyewitness identifications and lay judgments of handwriting evidence (Hasel & Kassin, 2009; Kukucka & Kassin, 2014), and the handwriting finding was replicated when the confession was reported by a jailhouse informant (Jenkins et al., 2021, as cited in Kassin et al., 2025). The 2025 scientific review also lists informants among the sources of evidence that knowledge of a confession can taint (Kassin et al., 2025). These studies support asking when an informant first reported the statement, what the informant knew about the case at that time and from what sources, and whether other evidence was evaluated after the informant’s account was known.

How Jurors Evaluate Secondary Confessions

In two mock-jury experiments, a secondary confession from an accomplice, a jailhouse informant, or a community member increased guilty verdicts compared with a no-confession control. Telling participants that the witness had received an incentive to testify, such as leniency or a reward, did not affect verdicts, and participants attributed the witness’s decision to testify mainly to personal rather than situational factors (Neuschatz et al., 2008). The authors interpreted this as an instance of the fundamental attribution error, the tendency to underweight situational pressures in explaining another person’s behavior. The 2025 scientific review, citing this study and later work (Wetmore et al., 2014), summarizes the research as showing that jurors do not fully discount a confession reported secondhand by an informant motivated to lie (Kassin et al., 2025).

The findings are not uniform. In a study of jury-eligible community members reading a murder transcript, the presence of an incentive did affect verdicts, ratings of the informant, and attributions, although the size of the incentive and expert testimony about the fundamental attribution error did not (Maeder & Pica, 2014). The authors suggested that potential jurors may have become more aware of incentive problems but remain insensitive to how large the benefit is. In two online experiments involving an alleged sexual assault of a 6-year-old or a 25-year-old, testimony from a jailhouse informant that the defendant had confessed made mock jurors’ judgments more favorable to the complainant, an effect that operated through increased ratings of the complainant’s credibility (Golding et al., 2023). That finding is relevant to sexual-offense prosecutions, in which complainant credibility is often the central question, and it suggests that informant testimony may affect how jurors weigh other witnesses as well as the accused.

These studies used written summaries or transcripts, individual mock jurors rather than deliberating juries, and online or student samples. They show that informant testimony can be persuasive and that disclosure of incentives does not reliably reduce its weight. They do not show how actual juries respond after full cross-examination, argument, and instructions. Juror research on confessions generally is summarized in “How Jurors Respond to Confession Evidence.”

Hypothetical Example. Consider a hypothetical shooting in a parking garage in which two jail inmates each tell investigators that the defendant described the crime to them. The first, who has asked for a sentence reduction, repeats details that appeared in news coverage and in the charging papers the defendant kept in his cell. The second, who has asked for nothing, says the defendant told him the gun was dropped into a storm drain behind a laundromat, a location in no public source; investigators search the drain and recover the weapon. The first account adds little that the informant could not have learned elsewhere, while the second contains information that led to evidence investigators did not have. The example illustrates that the research and the law direct attention to the source of each detail and to independent corroboration, not to the label “informant.”

Co-Accused Statements and Implicating Others

A co-accused who describes the crime may implicate the accused while minimizing his own role, and a co-accused who has been offered a plea agreement in exchange for testimony has a direct incentive. Accomplice testimony was one of the conditions in the jury study described above, and it increased convictions in the same way as informant testimony (Neuschatz et al., 2008). Accomplices often have firsthand knowledge of the offense, which is why their testimony can be both valuable and difficult to evaluate.

Documented false-confession cases show a related problem. Of 40 people exonerated by DNA testing who had falsely confessed, 17 also falsely implicated other people, and 11 of those others were later exonerated as well (Garrett, 2010). Where one person’s confession names others, the reliability of that naming depends on the same questions as any confession: how it was obtained, what details it contains and where they came from, and what independent evidence supports it. One co-accused’s confession is not independent corroboration of another’s if both statements were shaped by the same interviews or the same disclosed information.

Limits of the Evidence

The research on secondary confessions is smaller and newer than the research on primary confessions. The exoneration figures come from selected sets of known wrongful convictions; they show that informant testimony has contributed to errors, not how often it is wrong. The production studies use students in low-stakes tasks or community members making hypothetical choices, and the juror studies use written materials, individual decisions, and online or student samples. No study compares the accuracy of informants with and without incentives in actual cases, and none measures how often informant information leads to evidence that is later verified. The studies reviewed here did not report sex or gender differences in willingness to provide a secondary confession or in juror responses; one predicted that women would be more favorable to complainants but its published summary reports no such result (Golding et al., 2023). Research can identify conditions that make an informant’s account more or less reliable; it cannot establish whether a particular informant is telling the truth.

Questions to Ask About This Research

When the jury studies (Neuschatz et al., 2008; Wetmore et al., 2014) are cited to argue that jurors cannot evaluate informants, ask who the participants were, whether they deliberated, and whether the materials included the cross-examination and instructions that a trial would; also ask how the later study finding that incentives did affect verdicts (Maeder & Pica, 2014) bears on the claim. When the production studies (Swanner et al., 2010; Jenkins et al., 2023) are cited, ask whether a computer-crash task or a hypothetical offer resembles the informant’s actual situation. When exoneration data are cited, ask whether they describe a rate or a set of known errors. When informant testimony is offered as reliable because it contains details of the crime, ask what the informant could have learned from the accused’s paperwork, the news, other inmates, or investigators. These questions are explained in “Reading Research Claims: A Guide for Attorneys” on this site.

Secondary Confessions and the Law

Federal and Military Standards

Once formal charges have been brought and the Sixth Amendment right to counsel has attached, the government may not deliberately elicit incriminating statements from the accused about those charges through an informant in the absence of counsel (Massiah v. United States, 1964). In United States v. Henry (1980), the Court applied that rule to a paid informant placed in the accused’s cell block who was told not to question him but who engaged him in conversation; the government had intentionally created a situation likely to induce incriminating statements. In Kuhlmann v. Wilson (1986), by contrast, an informant who only listened, without deliberate action to elicit the statements, did not violate the right. Before charges, Miranda does not require warnings when an undercover agent posing as a fellow inmate questions an incarcerated suspect, because the coercion Miranda addresses is absent when the suspect does not know he is speaking to the police (Illinois v. Perkins, 1990). The prosecution must disclose promises and benefits given to a witness, because they bear on credibility (Giglio v. United States, 1972).

Statements by a co-accused raise confrontation issues. A non-testifying co-defendant’s confession that expressly implicates the defendant may not be admitted at a joint trial, even with a limiting instruction (Bruton v. United States, 1968); a confession redacted to remove the defendant’s name and any reference to his existence may be admitted with a limiting instruction (Richardson v. Marsh, 1987), but obvious redactions such as a blank or the word “deleted” do not suffice (Gray v. Maryland, 1998). The Court has since held that a confession altered to refer to “the other person,” with a limiting instruction, may be admitted, a holding about neutral references of that kind that does not reach the obvious redactions Gray addressed (Samia v. United States, 2023). Testimonial statements of a witness who does not testify are admissible against the accused only if the witness is unavailable and the accused had a prior opportunity to cross-examine (Crawford v. Washington, 2004). An accused’s admission reported by an informant is still a confession or admission subject to the corroboration requirement (Opper v. United States, 1954).

In courts-martial, Article 31(b) and Mil. R. Evid. 305(c) require warnings when a person subject to the Code interrogates or requests a statement from a suspect, and Rule 305(b)(1) extends the requirement to “a knowing agent” of such a person or of a military unit. Whether a questioner was acting in an official law-enforcement or disciplinary capacity is decided objectively and on the facts (United States v. Jones, 2014). After preferral, Rule 305(c)(3) bars use of statements obtained by anyone acting in a law-enforcement capacity, or that person’s agent, on the preferred charges if the accused has requested or has counsel. Rule 306 provides that at a joint trial a statement admissible only against one accused may not be received unless all references inculpating another accused are effectively deleted or the maker is subject to cross-examination. An informant-reported admission is subject to Rule 304(c) corroboration, and the members weigh it under Rule 304(g) and the ordinary credibility instructions.

Arizona

The Sixth Amendment, Miranda, Giglio, Bruton, and Crawford rules apply in Arizona prosecutions, and Arizona courts also apply the state constitutional right to counsel (Ariz. Const. art. 2, § 24). An admission reported by an informant must satisfy Arizona’s corpus delicti rule before it may be considered (State v. Carlson, 2015). As of November 2026, we located no informant-specific jury instruction in Arizona’s standard criminal instructions; jurors are told to consider, among other things, whether a witness has any motive, bias, or prejudice and whether the witness is contradicted by prior statements or other evidence (RAJI (Criminal) Preliminary 10, 2022). Some states have adopted informant-specific procedures; Illinois, for example, requires disclosure of an in-custody informant’s criminal history, benefits, prior testimony, and any recantation in specified homicide, sexual-assault, and arson prosecutions and requires the prosecution to prove the testimony reliable by a preponderance at a pretrial hearing (725 ILCS 5/115-21). We located no comparable Arizona statute; current rules and statutes should be verified before use.

Special Issues in Legal Practice

The research described above is the same whichever party relies on it. The science operates the same way regardless of who is asking the questions, and the research summarized here does not by itself favor either side. What differs between the parties is not the science but the vantage point — the particular places where these issues tend to surface, and the particular errors each party is best positioned to guard against. The two summaries below identify those recurring issues. They are offered for orientation, not as litigation strategy; in every case the controlling question is what the evidence in that case will, and will not, support, and the expert’s role is to inform the trier of fact rather than to advocate for either side.

Special Issues for Prosecutors

The research does not suggest that informant or co-accused testimony is generally false, and prosecutors and investigators report relying on cooperating witnesses with firsthand knowledge in serious cases. The research supports the practices that make such testimony most reliable: corroborating the account with facts the witness could not have learned from the accused’s papers, media, other inmates, or investigators; documenting when and how the witness first came forward and what he or she knew about the case at that time; recording the witness’s statements to investigators; and disclosing all benefits fully, as Giglio requires. Information from an informant that leads to evidence not previously known to investigators, and that the informant could not have learned from other sources, is the strongest corroboration. Courts have held that statements deliberately elicited by informants acting at the government’s direction after charges are subject to exclusion under Massiah and Henry, and in courts-martial, whether a service-member informant is a knowing agent under Rule 305(b)(1) bears on admissibility under Article 31.

The research also counsels caution. Incentives increased false secondary reports in the laboratory (Swanner et al., 2010; Jenkins et al., 2023), in some mock-juror studies participants did not discount informant testimony when told of incentives (Neuschatz et al., 2008; Wetmore et al., 2014), although one study found that incentives did affect verdicts (Maeder & Pica, 2014), and informant testimony appears among the errors in known wrongful convictions, often following a confession (Kassin et al., 2012). In sexual-offense cases, informant testimony may raise jurors’ assessment of the complainant’s credibility as well as the case against the accused (Golding et al., 2023), which makes its reliability more consequential. Co-accused statements implicating others warrant the same scrutiny as any confession (Garrett, 2010).

Special Issues for Defense

The research supports close examination of an informant’s or co-accused’s account. Defense counsel are often best positioned to establish what the witness could have learned from sources other than the accused, such as discovery papers in the accused’s possession, news coverage, other inmates, or the questions investigators asked; when the witness first came forward; what benefits were sought, offered, or expected; whether the witness has testified in other cases; and whether the account changed over time. Laboratory findings on incentives (Swanner et al., 2010; Jenkins et al., 2023) and on the persuasiveness of secondary confessions (Neuschatz et al., 2008; Wetmore et al., 2014) bear on cross-examination, requested instructions, and any expert testimony the court allows, and Massiah, Henry, Article 31, and Rule 305 bear on whether the statement is admissible at all.

The research also counsels caution. No study shows that informants in general, or a particular informant, are lying, and some informants report accurately what an accused told them. A co-accused with firsthand knowledge may give detailed testimony that is corroborated by physical or digital evidence, and in that circumstance general research has little to say about the account. One study found that jurors did take incentives into account (Maeder & Pica, 2014), and expert testimony on attribution error did not change verdicts in that study, so the value of such testimony is uncertain. The research is most informative when directed at the specific details the witness could have learned elsewhere or at an incentive the witness concealed, rather than at informant testimony as a category.

This paper is for educational purposes only and is not legal advice. The author is a psychologist, not an attorney; summaries of law are general and must be verified against current authority or with a licensed attorney.

References

Garrett, B. L. (2010). The substance of false confessions. Stanford Law Review, 62(4), 1051–1118.

Golding, J. M., Neuschatz, J. S., Bornstein, B. H., Pals, A. M., & Wetmore, S. A. (2023). The perception of a jailhouse informant in a sexual assault case. Journal of Police and Criminal Psychology, 38(2), 281–292. https://doi.org/10.1007/s11896-020-09400-3

Hasel, L. E., & Kassin, S. M. (2009). On the presumption of evidentiary independence: Can confessions corrupt eyewitness identifications? Psychological Science, 20(1), 122–126. https://doi.org/10.1111/j.1467-9280.2008.02262.x

Jenkins, B. D., Le Grand, A. M., Wetmore, S. A., Neuschatz, J. S., Golding, J. M., & Lippert, A. (2023). A snitching enterprise: The role of evidence and incentives on providing false secondary confessions. Journal of Police and Criminal Psychology, 38(1), 141–152. https://doi.org/10.1007/s11896-022-09502-0

Kassin, S. M., Bogart, D., & Kerner, J. (2012). Confessions that corrupt: Evidence from the DNA exoneration case files. Psychological Science, 23(1), 41–45. https://doi.org/10.1177/0956797611422918

Kassin, S. M., Cleary, H. M. D., Gudjonsson, G. H., Leo, R. A., Meissner, C. A., Redlich, A. D., & Scherr, K. C. (2025). Police-induced confessions, 2.0: Risk factors and recommendations. Law and Human Behavior, 49(1), 7–53. https://doi.org/10.1037/lhb0000593

Kukucka, J., & Kassin, S. M. (2014). Do confessions taint perceptions of handwriting evidence? An empirical test of the forensic confirmation bias. Law and Human Behavior, 38(3), 256–270. https://doi.org/10.1037/lhb0000066

Maeder, E. M., & Pica, E. (2014). Secondary confessions: The influence (or lack thereof) of incentive size and scientific expert testimony on jurors’ perceptions of informant testimony. Law and Human Behavior, 38(6), 560–568. https://doi.org/10.1037/lhb0000106

Neuschatz, J. S., Lawson, D. S., Swanner, J. K., Meissner, C. A., & Neuschatz, J. S. (2008). The effects of accomplice witnesses and jailhouse informants on jury decision making. Law and Human Behavior, 32(2), 137–149. https://doi.org/10.1007/s10979-007-9100-1

Swanner, J. K., Beike, D. R., & Cole, A. T. (2010). Snitching, lies and computer crashes: An experimental investigation of secondary confessions. Law and Human Behavior, 34(1), 53–65. https://doi.org/10.1007/s10979-008-9173-5

Warden, R. (2004). The snitch system: How snitch testimony sent Randy Steidl and other innocent Americans to death row. Northwestern University School of Law, Center on Wrongful Convictions.

Wetmore, S. A., Neuschatz, J. S., & Gronlund, S. D. (2014). On the power of secondary confession evidence. Psychology, Crime & Law, 20(4), 339–357. https://doi.org/10.1080/1068316X.2013.777963

Legal Authorities

Bruton v. United States, 391 U.S. 123 (1968).

Crawford v. Washington, 541 U.S. 36 (2004).

Giglio v. United States, 405 U.S. 150 (1972).

Gray v. Maryland, 523 U.S. 185 (1998).

Illinois v. Perkins, 496 U.S. 292 (1990).

Kuhlmann v. Wilson, 477 U.S. 436 (1986).

Massiah v. United States, 377 U.S. 201 (1964).

Opper v. United States, 348 U.S. 84 (1954).

Richardson v. Marsh, 481 U.S. 200 (1987).

Samia v. United States, 599 U.S. 635 (2023).

United States v. Henry, 447 U.S. 264 (1980).

State v. Carlson, 237 Ariz. 381, 351 P.3d 1079 (2015).

United States v. Jones, 73 M.J. 357 (C.A.A.F. 2014).

U.S. Const. amends. V, VI.

Ariz. Const. art. 2, § 24.

Uniform Code of Military Justice art. 31, 10 U.S.C. § 831.

725 Ill. Comp. Stat. 5/115-21.

Manual for Courts-Martial, United States, Mil. R. Evid. 304, 305, 306 (2024 ed.).

Revised Arizona Jury Instructions (Criminal), Preliminary Criminal 10 (State Bar of Arizona 2022).

Related topic papers: Corroboration and Assessing the Reliability of a Confession · Contamination: Non-Public Details and Post-Admission Narratives · Investigator Bias, Tunnel Vision, and Forensic Confirmation Bias · How Jurors Respond to Confession Evidence · Confessions in Courts-Martial: Article 31(b), Mil. R. Evid. 304 and 305, and Expert Assistance

Related topics from Arizona Forensic Services: Jurors and Jury Decision-Making · Courts-Martial