Snap Shot
- The question: How can an investigator’s early belief in a suspect’s guilt shape the interrogation, the interpretation of other evidence, and the evidence itself, and what safeguards reduce that risk?
- What the research shows: In laboratory experiments, mostly with students, interviewers led to expect guilt asked more guilt-presumptive questions and applied more pressure, especially to innocent suspects, and investigator bias increased minimization and false confessions without affecting guilty participants. Knowledge of a confession changed eyewitness identifications and lay judgments of forensic evidence, and context changed some fingerprint examiners’ conclusions.
- What it does not show: Most of this research is laboratory-based, often with students or trainees; field evidence is largely archival and correlational. The research does not show that investigators generally act in bad faith, that forming a hypothesis is improper, or that bias affected any particular case.
- The law: The prosecution must disclose favorable evidence, including information that would allow the defense to attack the thoroughness and good faith of an investigation (Brady; Kyles v. Whitley; R.C.M. 701(a)(6); Ariz. R. Crim. P. 15.1(b)). Arizona’s Willits instruction addresses lost material evidence; a less thorough investigation alone does not require it.
- Issues for each side: For prosecutors, the research favors safeguards that keep forensic and witness evidence independent of the confession, which protects that evidence; for defense counsel, identifying specific points where an early conclusion shaped what was collected or how it was interpreted, rather than general claims of tunnel vision.
Every investigation requires judgment about who is likely responsible. Investigators form hypotheses early, pursue leads, and decide where to spend limited time; without that, few cases would be solved. The risk researchers describe is not that investigators form hypotheses but that a hypothesis, once formed, may narrow what is looked for, how ambiguous information is read, and how contrary information is weighed. In wrongful-conviction research, this pattern is called tunnel vision, and its psychological basis is confirmation bias, the general human tendency to seek and interpret information in ways that fit existing beliefs (Findley, 2012; Leo, 2008).
These issues reach court in several ways. A defendant may argue that an interrogation was guilt-presumptive from the start, that alternative suspects were not pursued, or that forensic analysts or eyewitnesses learned of a confession before reaching their conclusions. The prosecution may need to show that its evidence was developed independently. Disclosure obligations, lost-evidence instructions, and expert testimony all turn in part on these questions.
This paper reviews the research on hypothesis-driven investigation, presumption of guilt in interviews, and the influence of a confession on other evidence, together with the safeguards that agencies and laboratories use. It then summarizes the law. Confirmation bias as a general memory and reasoning phenomenon is addressed in the Memory in the Courtroom paper Confirmation Bias, and its role in child abuse investigations in the Child Abuse Allegations resource.
Hypotheses and Tunnel Vision in Investigations
Investigative doctrine itself calls for hypotheses. An Army law-enforcement manual directs investigators to develop a working theory of the offense and to reassess it as facts emerge, and it cautions against forcing facts that contradict the theory into existing assumptions (Department of the Army, 2005). One former homicide detective writes that theories formed before the facts are in are often partly wrong, and that once a theory hardens, investigators may stop testing it (Trainum, 2016); this is a first-hand account, not a measured rate. Findley (2012), reviewing wrongful-conviction research, describes tunnel vision as arising from ordinary cognitive tendencies reinforced by institutional pressures such as caseloads, public attention, and the adversarial structure itself.
Experimental work supports parts of this account and qualifies others. In three studies in which participants read a case file and chose further investigative steps, choices were guided by their initial assessment of guilt, and the tendency to choose incriminating steps rose with crime severity and the strength of the existing evidence; yet overall, participants did not favor incriminating over exonerating steps, which hardly exceeded half of the choices (Rassin et al., 2010). In a study with police trainees, evidence that disconfirmed their suspicion was rated less reliable than identical evidence that confirmed it, an asymmetry that was stronger for witness evidence than for DNA or photographs (Ask et al., 2008). These findings suggest that bias appears most in how flexible evidence is interpreted, and that its strength varies with circumstances.
Presumption of Guilt in the Interview Room
The decision to interrogate usually reflects a belief that the suspect is guilty, and that belief can shape the interview. In a two-phase experiment, student interrogators led to expect guilt chose more guilt-presumptive questions, used more techniques, and exerted more pressure, especially with innocent suspects; neutral listeners then judged suspects in that condition as more defensive and somewhat more guilty (Kassin et al., 2003). Using a different paradigm, Narchet et al. (2011) found that investigator bias increased the use of minimization tactics and thereby increased false confessions by innocent participants, while it had no direct or indirect effect on guilty participants.
Training and experience can shift judgment toward suspicion without improving accuracy. In one laboratory study of North American law-enforcement investigators judging recorded truthful and deceptive accounts, training and experience did not improve discrimination but increased the tendency to judge accounts as deceptive (Meissner & Kassin, 2002). That pattern was termed investigator bias; in that study, confidence in a judgment of deception was not a reliable indicator of its accuracy. Group findings from one laboratory task do not show how any particular investigator performs. Deception detection is addressed in The Pre-Interrogation Interview: Behavior Analysis, Investigator Judgment, and the Polygraph and in the Memory hub’s Detecting Deception.
Interview methods matter here. Information-gathering approaches that begin with the suspect’s own account and test it against evidence are designed in part to keep the interviewer’s hypothesis open; they are discussed in Accusatorial and Information-Gathering Methods: What the Comparative Research Shows. In a mock crime-and-investigation study with 61 investigators interviewing mock suspects, those who knew the session was recorded were better at distinguishing guilty from innocent suspects (Kassin et al., 2014), suggesting that accountability may itself reduce some effects of expectation.
When a Confession Shapes Other Evidence
A confession can influence the evidence that is supposed to test it. Kassin, Dror, and Kukucka (2013) used the term forensic confirmation bias for the ways in which a suspect’s confession, or other contextual information, can affect the collection, perception, and interpretation of forensic and other evidence. In a laboratory study, many participants who had chosen someone from a lineup that did not contain the perpetrator changed their identification after being told that another lineup member had confessed (Hasel & Kassin, 2009). Lay participants told that a defendant had confessed were more likely to judge handwriting samples as matching (Kukucka & Kassin, 2014). Participants who had corroborated a person’s alibi were much less likely to maintain it after learning she had confessed (Marion et al., 2016); see Alibis and the Innocent Suspect.
Forensic experts are not immune to context. In a very small study, five latent print examiners who had previously identified prints as matching were later shown the same prints with information suggesting the suspect could not be the source, and four changed their conclusions (three to no match, one to no definite decision) (Dror et al., 2006). A systematic review of 29 studies of cognitive bias in forensic science found evidence of confirmation bias in most studies of practitioners given case-specific information about the suspect or crime, though the evidence was concentrated in fingerprint analysis, with only a few studies in other disciplines (Cooper & Meterko, 2019). In a survey of 403 forensic examiners in 21 countries, many regarded bias as a concern in forensic science generally but were less likely to see it in their own judgments, and many believed they could set it aside by willpower (Kukucka et al., 2017).
Archival data point in the same direction. In DNA exoneration cases, false-confession cases were more likely than eyewitness cases to contain other errors, including improper forensic science, mistaken identifications, and informant testimony, and the confession was usually obtained first (Kassin et al., 2012). The archival data are consistent with the laboratory findings but cannot show that the confession caused the later errors in any case.
The research has critics. A narrative review in the Journal of Forensic Sciences argued that methodological limitations and questions of ecological validity restrict how far laboratory findings on contextual bias apply to operational casework, and that task-irrelevant context does not always produce error; the authors called for joint studies by forensic scientists and psychologists in realistic settings (Curley et al., 2020). Those cautions are well taken: the size of these effects in real laboratories and investigations is not known.
Case Example. According to the National Registry of Exonerations, 16-year-old Jeffrey Deskovic was questioned in January 1990 about the 1989 rape and murder of a 15-year-old classmate in Peekskill, New York, in sessions that included polygraph examinations and were not all recorded; police said he confessed. Before trial, DNA testing excluded him as the source of semen from the victim. The prosecution proceeded on the confession, arguing that the victim had consensual sex with someone else and that Deskovic killed her out of jealousy, and a jury convicted him in 1990. In 2006, retesting matched the DNA profile to a man already imprisoned for another murder, who later pleaded guilty, and the charges against Deskovic were dismissed (National Registry of Exonerations, n.d.). The case illustrates how a confession can lead investigators and factfinders to reinterpret exculpatory evidence rather than revisit the hypothesis.
Safeguards
Researchers and practitioners have proposed procedures that keep evidence independent of conclusions already reached. In forensic science, Linear Sequential Unmasking and its expanded form ask analysts to examine the evidence before seeing reference samples or task-irrelevant case information, to document conclusions at each stage, and to receive further information only as needed (Dror & Kukucka, 2021); a practical worksheet has been published to help laboratories apply the approach (Quigley-McBride et al., 2022). For eyewitness procedures, double-blind lineup administration serves a similar purpose. The 2025 scientific review recommends shielding forensic examiners and witnesses from knowledge of a confession where possible (Kassin et al., 2025).
Within investigations, proposed safeguards include recording key decisions and the reasons for them, revisiting those decisions when new evidence arrives, assigning a reviewer to argue the alternative, and supervisory or peer review of major cases (Trainum, 2016; Findley, 2012). Complete recording of interrogations allows later reviewers to see whether questioning remained open to the suspect’s account. These safeguards have been evaluated less thoroughly than the biases they address, and their effects in operational settings are still being studied.
Hypothetical Example. Consider a hypothetical burglary in which a suspect confesses and the case file is sent to a laboratory for comparison of a shoe impression. Under the laboratory’s sequential procedure, the examiner documents the features of the crime-scene impression before seeing the suspect’s shoe and is not told of the confession. The examiner reports a correspondence in class characteristics and two specific wear features. At trial, the defense argues that the forensic evidence was shaped by the confession; the documentation shows the examiner’s observations were recorded before any comparison and without that knowledge. The example illustrates that independence safeguards protect sound forensic evidence for the prosecution as much as they guard against error.
Limits of the Evidence
The experimental evidence on investigator bias comes mostly from laboratory studies with students, police trainees, or small groups of practitioners, using short case files and mock interrogations. Field evidence is largely archival, drawn from exoneration cases, and cannot establish causation or frequency. The forensic bias literature is concentrated in a few disciplines, and critics have questioned how well its designs reflect casework (Cooper & Meterko, 2019; Curley et al., 2020). Effect sizes vary, and some studies find that participants do not, on average, favor incriminating information (Rassin et al., 2010).
Few of these studies analyzed sex or gender. The interrogation experiments used mixed student samples without reported sex differences, the mock suspect in one field study was male, and the forensic surveys and studies did not report differences by examiner sex in the sources reviewed. Group research can show that expectations can shape investigations; it cannot show that bias affected a particular investigator, interview, or analysis.
Questions to Ask About This Research
When bias research is cited, ask whether the participants were students, trainees, or working professionals, whether the task resembled real casework, and whether the study measured actual errors or only shifts in judgment. For a particular case, ask when the suspect became the focus, what leads were pursued after that point, whether forensic analysts or witnesses knew of the confession before reaching their conclusions, whether analysts documented their observations before comparison, and whether any review process was used. These questions are explained in “Reading Research Claims: A Guide for Attorneys” on this site.
Investigator Bias and the Law
Federal and Military Standards
Due process requires the prosecution to disclose evidence favorable to the accused that is material to guilt or punishment (Brady v. Maryland, 1963). In Kyles v. Whitley (1995), the Supreme Court held that materiality is assessed by the cumulative effect of all suppressed evidence, that the prosecutor is responsible for favorable evidence known to the police, and that undisclosed information can be material because it would have allowed the defense to challenge the thoroughness and even the good faith of the investigation. The failure to preserve potentially useful evidence violates due process only on a showing of bad faith (Arizona v. Youngblood, 1988). Expert testimony about bias research is governed by Fed. R. Evid. 702, as amended Dec. 1, 2023, and Kumho Tire Co. v. Carmichael (1999).
In courts-martial, trial counsel must disclose, as soon as practicable, evidence known to trial counsel that tends to negate guilt, reduce the degree of guilt or the punishment, or adversely affect the credibility of prosecution witnesses or evidence (R.C.M. 701(a)(6), Manual for Courts-Martial, 2024 ed.). Mil. R. Evid. 702 governs expert testimony; the 2023 federal amendment to Rule 702 applies to the military rule from June 1, 2025, 18 months after its December 1, 2023 effective date, under Mil. R. Evid. 1102(a) unless the President acts otherwise (Joint Service Committee on Military Justice, 2025).
Arizona
Arizona requires the State to disclose material or information that tends to mitigate or negate the defendant’s guilt (Ariz. R. Crim. P. 15.1(b)). When the State fails to preserve material evidence that was reasonably accessible and could have had a tendency to exonerate the accused, and prejudice results, the defendant is entitled to an adverse-inference instruction under State v. Willits (1964), as clarified in State v. Glissendorf (2014) and State v. Hernandez (2020). The instruction is not required merely because a more thorough investigation could have been made (State v. Murray, 1995). Experts may describe research on bias but may not offer opinions on a particular witness’s truthfulness (State v. Lindsey, 1986, 149 Ariz. at 475).
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 supports the prosecution’s use of evidence developed under procedures that keep it independent of the confession. Forensic conclusions documented before the analyst knew of a confession, lineups administered by someone who did not know the suspect, and investigative records showing that alternative leads were pursued all answer the concerns the research raises (Dror & Kukucka, 2021; Kassin et al., 2025). Forming a hypothesis is a necessary part of investigation (Department of the Army, 2005), and studies show that bias effects vary and that investigators do not uniformly favor incriminating information (Rassin et al., 2010). Critiques of laboratory designs (Curley et al., 2020) are legitimate points about how far the research generalizes.
The research also counsels caution. Evidence that an examiner or witness learned of a confession before reaching a conclusion raises a question about independence that the research can inform (Hasel & Kassin, 2009; Kassin et al., 2012), and its weight is for the court; Kyles makes clear that information bearing on the thoroughness of the investigation may be subject to disclosure. When exculpatory evidence is explained away rather than investigated, the explanation may later be scrutinized. Contemporaneous documentation is easier to evaluate than explanations offered later.
Special Issues for Defense
The research supports specific inquiry into the sequence of an investigation: when the suspect became the focus, what was done afterward, and whether forensic analysts and witnesses knew of the confession (Kassin et al., 2013; Cooper & Meterko, 2019). Defense counsel are often best positioned to identify leads that were not pursued and evidence that was reinterpreted to fit the theory, and Kyles supports seeking disclosure of information bearing on the investigation’s thoroughness. Experimental findings on guilt-presumptive questioning (Kassin et al., 2003; Narchet et al., 2011) can help explain how an interview was conducted when the recording shows that pattern.
The research also counsels caution. Most of it is laboratory-based, its effect sizes in real casework are not known, and critics have identified real limitations (Curley et al., 2020). Focusing on a suspect is not itself evidence of bias, and a claim that an investigation was biased has little support without specific points where the hypothesis shaped what was collected or concluded. Where forensic evidence was developed blind to the confession, the forensic confirmation bias research does not apply to it.
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
Ask, K., Rebelius, A., & Granhag, P. A. (2008). The ‘elasticity’ of criminal evidence: A moderator of investigator bias. Applied Cognitive Psychology, 22(9), 1245–1259. https://doi.org/10.1002/acp.1432
Cooper, G. S., & Meterko, V. (2019). Cognitive bias research in forensic science: A systematic review. Forensic Science International, 297, 35–46. https://doi.org/10.1016/j.forsciint.2019.01.016
Curley, L. J., Munro, J., Lages, M., MacLean, R., & Murray, J. (2020). Assessing cognitive bias in forensic decisions: A review and outlook. Journal of Forensic Sciences, 65(2), 354–360. https://doi.org/10.1111/1556-4029.14220
Department of the Army. (2005). Law enforcement investigations (Field Manual 3-19.13).
Dror, I. E., Charlton, D., & Péron, A. E. (2006). Contextual information renders experts vulnerable to making erroneous identifications. Forensic Science International, 156(1), 74–78. https://doi.org/10.1016/j.forsciint.2005.10.017
Dror, I. E., & Kukucka, J. (2021). Linear Sequential Unmasking–Expanded (LSU-E): A general approach for improving decision making as well as minimizing noise and bias. Forensic Science International: Synergy, 3, 100161. https://doi.org/10.1016/j.fsisyn.2021.100161
Findley, K. A. (2012). Tunnel vision. In B. L. Cutler (Ed.), Conviction of the innocent: Lessons from psychological research (pp. 303–323). American Psychological Association. https://doi.org/10.1037/13085-014
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
Joint Service Committee on Military Justice. (2025). Amendments to the Military Rules of Evidence since 1984 (G. E. Maggs, Ed.; 2025 ed.). https://jsc.defense.gov/Portals/99/MRE%20Amendments%202025.pdf
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
Kassin, S. M., Dror, I. E., & Kukucka, J. (2013). The forensic confirmation bias: Problems, perspectives, and proposed solutions. Journal of Applied Research in Memory and Cognition, 2(1), 42–52. https://doi.org/10.1016/j.jarmac.2013.01.001
Kassin, S. M., Goldstein, C. C., & Savitsky, K. (2003). Behavioral confirmation in the interrogation room: On the dangers of presuming guilt. Law and Human Behavior, 27(2), 187–203. https://doi.org/10.1023/A:1022599230598
Kassin, S. M., Kukucka, J., Lawson, V. Z., & DeCarlo, J. (2014). Does video recording alter the behavior of police during interrogation? A mock crime-and-investigation study. Law and Human Behavior, 38(1), 73–83. https://doi.org/10.1037/lhb0000047
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
Kukucka, J., Kassin, S. M., Zapf, P. A., & Dror, I. E. (2017). Cognitive bias and blindness: A global survey of forensic science examiners. Journal of Applied Research in Memory and Cognition, 6(4), 452–459. https://doi.org/10.1016/j.jarmac.2017.09.001
Leo, R. A. (2008). Police interrogation and American justice. Harvard University Press.
Marion, S. B., Kukucka, J., Collins, C., Kassin, S. M., & Burke, T. M. (2016). Lost proof of innocence: The impact of confessions on alibi witnesses. Law and Human Behavior, 40(1), 65–71. https://doi.org/10.1037/lhb0000156
Meissner, C. A., & Kassin, S. M. (2002). “He’s guilty!”: Investigator bias in judgments of truth and deception. Law and Human Behavior, 26(5), 469–480. https://doi.org/10.1023/A:1020278620751
Narchet, F. M., Meissner, C. A., & Russano, M. B. (2011). Modeling the influence of investigator bias on the elicitation of true and false confessions. Law and Human Behavior, 35(6), 452–465. https://doi.org/10.1007/s10979-010-9257-x
National Registry of Exonerations. (n.d.). Jeffrey Deskovic. Retrieved November 5, 2026, from https://exonerationregistry.org/cases/10405
Quigley-McBride, A., Dror, I. E., Roy, T., Garrett, B. L., & Kukucka, J. (2022). A practical tool for information management in forensic decisions: Using Linear Sequential Unmasking-Expanded (LSU-E) in casework. Forensic Science International: Synergy, 4, 100216. https://doi.org/10.1016/j.fsisyn.2022.100216
Rassin, E., Eerland, A., & Kuijpers, I. (2010). Let’s find the evidence: An analogue study of confirmation bias in criminal investigations. Journal of Investigative Psychology and Offender Profiling, 7(3), 231–246. https://doi.org/10.1002/jip.126
Trainum, J. L. (2016). How the police generate false confessions: An inside look at the interrogation room. Rowman & Littlefield.
Legal Authorities
Arizona v. Youngblood, 488 U.S. 51 (1988).
Brady v. Maryland, 373 U.S. 83 (1963).
Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999).
Kyles v. Whitley, 514 U.S. 419 (1995).
State v. Glissendorf, 235 Ariz. 147, 329 P.3d 1049 (2014).
State v. Hernandez, 250 Ariz. 28 (2020).
State v. Lindsey, 149 Ariz. 472, 720 P.2d 73 (1986).
State v. Murray, 184 Ariz. 9 (1995).
State v. Willits, 96 Ariz. 184, 393 P.2d 274 (1964).
Fed. R. Evid. 702 (as amended Dec. 1, 2023).
Ariz. R. Crim. P. 15.1.
Manual for Courts-Martial, United States, R.C.M. 701(a)(6) (2024 ed.).
Manual for Courts-Martial, United States, Mil. R. Evid. 702, 1102 (2024 ed.; Mil. R. Evid. 702 as amended 2025).
Related topic papers: Corroboration and Assessing the Reliability of a Confession · Contamination: Non-Public Details and Post-Admission Narratives · Alibis and the Innocent Suspect · The Pre-Interrogation Interview: Behavior Analysis, Investigator Judgment, and the Polygraph · Accusatorial and Information-Gathering Methods: What the Comparative Research Shows · Secondary Confessions: Informants, Co-Accused, and Jailhouse Statements
Related topics from Arizona Forensic Services: Memory in the Courtroom · Child Abuse Allegations