Snap Shot
- The question: Can a belief formed early in a case shape what investigators look for, what interviewers ask, how examiners read evidence, how jurors weigh it, and what witnesses come to remember?
- What the research shows: People tend to seek, interpret, and recall information in ways that fit an existing belief. Experiments have found this in police officers, interviewers, forensic examiners, forensic mental health experts, and mock jurors, and interviewers’ expectations can shift what witnesses report. Blind and sequential procedures are designed to limit it.
- What it does not show: It does not show that any particular investigation, interview, or examination was biased or reached a wrong result. Effects vary across tasks and studies, at least one well-designed study with trained examiners found no measurable effect, and many studies use students and simulated materials.
- The law: Rule 702 and Houser govern expert testimony on bias, which stays general and may not declare a particular person biased or a witness untruthful (Lindsey). Evidence that bears on the reliability of an investigation can be favorable evidence under Brady and Kyles and R.C.M. 701(a)(6); Arizona’s Willits instruction addresses evidence the State failed to preserve and is not required merely because an investigation could have been more thorough.
- Issues for each side: Prosecutors can strengthen their evidence by documenting what each investigator, interviewer, and examiner knew and when, and by using blind procedures; defense counsel can examine the sequence and independence of the evidence, recognizing that the same bias operates on the defense.
Confirmation bias is the tendency to look for, interpret, and remember information in ways that support what a person already believes or expects, and to give less weight to information that points the other way. It is among the most thoroughly documented tendencies in research on human reasoning. It is not dishonesty: it operates largely outside awareness, and people affected by it usually believe they are weighing the evidence fairly.
Confirmation bias reaches court because cases are built in sequence. An early theory of who is responsible can influence which leads are pursued, what witnesses are asked, what forensic examiners are told, and how ambiguous evidence is described. By the time a case is tried, several items of evidence may appear to confirm one another when some of them were influenced by the same early belief. The same processes affect defense investigators and experts, counsel on both sides, and jurors.
This paper summarizes the research on confirmation bias in investigations, interviews, forensic examinations, expert evaluations, and juror reasoning; the proposed safeguards; the limits of the evidence; and the federal, military, and Arizona law the topic touches.
What Confirmation Bias Is
In a widely cited review, Nickerson (1998) described confirmation bias as a family of related tendencies rather than a single error: selective search for supporting information, interpretation of ambiguous information to fit an expectation, and selective recall of confirming over disconfirming information. A related phenomenon is belief perseverance, in which a belief survives after its basis has been discredited.
Klayman and Ha (1987) refined the classic account. They argued that much of what is called confirmation bias reflects a “positive test strategy”: people test a hypothesis by examining cases in which they expect it to hold. That strategy is often reasonable, but it can produce systematic error when a hypothesis can be shown wrong only by examining cases in which it is not expected to hold. A theory tested only by looking for evidence that would be present if it were true may never meet the evidence that would show it false.
Expectation can also shape how the same evidence is perceived. In a classic experiment, participants watched an identical video of a child taking an academic test; those led to believe the child came from an advantaged background rated her ability higher than those led to believe she came from a disadvantaged background, and both groups cited features of the same performance in support of their ratings (Darley & Gross, 1983). The study is dated and concerns social judgment, but the same pattern recurs in the forensic research below.
Investigator Tunnel Vision
In the legal literature, the case-level form of confirmation bias is often called tunnel vision: once investigators settle on a suspect, the investigation may focus on building a case against that person while evidence pointing elsewhere is discounted or not pursued (Findley & Scott, 2006). They described it as ordinary cognition reinforced by institutional pressures, not bad faith, and as a recurring feature of wrongful convictions.
Experimental studies have tested parts of this account. Participants who were asked to name a suspect partway through reviewing an investigative file later showed more bias toward that suspect in how they evaluated and pursued evidence than participants who were not asked to commit, and asking participants to consider reasons why their favored suspect might be innocent reduced the effect (O’Brien, 2009). In a study with police officers, officers’ evaluations of ambiguous evidence tended to be consistent with what they already believed about the suspect’s guilt (Charman et al., 2017). Both used case materials rather than live investigations.
Tunnel vision also bears on how evidence accumulates. Kassin et al. (2013) described how a confession, an identification, or a forensic conclusion obtained with knowledge of the others can appear to corroborate them while not being independent of them. Independently produced evidence that converges carries more weight, and the record of how each item was obtained is often the only way to tell the two apart.
Interviewer Bias and Witness Memory
Confirmation bias affects the questions people ask, and questions can affect what witnesses report. Accuracy-focused retelling can strengthen recall; slanted retelling, repeated suggestive questioning, or exposure to others’ accounts can contaminate it (see “Remembering Changes Memory”). In experiments with student interviewers, those led to expect guilt asked more guilt-presumptive questions, and listeners judged suspects questioned that way as more defensive or more guilty, whether or not the suspect had committed the act (Hill et al., 2008; Kassin et al., 2003). Interrogators with guilty expectations also used more pressure, particularly with innocent suspects (Kassin et al., 2003), although in one study questioning style was not associated with whether suspects confessed (Hill et al., 2008).
With child witnesses, the interviewer’s belief can enter the child’s account. In a frequently cited experiment, young children watched a man presented as a janitor either clean or play with dolls; when an interviewer questioned them in a way that suggested an incriminating or an exculpating interpretation, the children’s accounts tended to shift toward that interpretation, and the shifts carried into later questioning by their parents, while neutral questioning produced accurate accounts (Thompson et al., 1997). The lesson cuts both ways: an interviewer holding either an incriminating or an exculpating hypothesis can shape a report, while open-ended questioning that tests more than one possibility is associated with accurate accounts.
Forensic Examiners and Expert Witnesses
Forensic comparison disciplines were once assumed to be largely immune to expectation effects. In a frequently cited study, five experienced fingerprint examiners were shown, without being told, pairs of prints they had each identified as matches in earlier casework, now accompanied by context suggesting the prints did not match; most gave judgments that departed from their own earlier identifications (Dror et al., 2006). Kassin et al. (2013) drew on this and related findings to describe a “forensic confirmation bias,” and a national review had earlier identified contextual bias as a concern requiring research and safeguards (National Research Council, 2009).
Later experiments extended the pattern to other tasks. Participants who learned that a defendant had confessed were more likely to conclude, mistakenly, that two handwriting samples came from the same writer (Kukucka & Kassin, 2014); the participants were laypeople rather than handwriting examiners. In a 2025 experiment simulating facial recognition searches, participants rated whichever candidate face was randomly paired with guilt-suggestive biographical information, or with a high software confidence score, as looking most like the perpetrator, and candidates paired with guilt-suggestive information were most often misidentified (Stewart & Kukucka, 2025).
The findings are not uniform. In a 2024 study comparing expert fingerprint examiners with students, varying the amount of case information given before the comparison, and redacting task-irrelevant information, had no measurable effect on decisions; the experts were more accurate than students on fingerprints but not on footwear impressions, outside their training (Pena et al., 2024). The fair summary is that contextual information can influence forensic judgments, particularly when the evidence is ambiguous, and that the size of the effect depends on the task, the examiner, and the procedure.
Forensic mental health experts, including those retained by the defense, are subject to the same processes. In an experiment in which 108 forensic psychologists and psychiatrists scored the same offender files on two risk-assessment instruments, those who believed they had been retained by the prosecution tended to assign higher risk scores and those who believed they had been retained by the defense tended to assign lower scores (Murrie et al., 2013). In a survey of forensic mental health evaluators, most agreed that cognitive bias is a concern in the field, but fewer believed their own judgments were affected, and most believed that evaluators who consciously set aside their expectations are less affected by them, a belief the authors questioned (Zapf et al., 2018). A systematic review of 23 studies in forensic mental health found significant or partial bias effects in most of the studies that tested for them, including adversarial allegiance, confirmation, and hindsight effects, and found comparatively little research on methods for reducing bias (Neal et al., 2022). These findings apply to any expert on either side, including memory experts.
Jurors and the Story Model
Jurors also bring expectations to evidence. According to the story model, jurors organize trial evidence into a narrative of what happened and choose the verdict that best fits the story they have accepted; in mock-juror experiments, evidence presented in an order that made one side’s story easier to construct favored that side, and confidence was related to the coherence of the story (Pennington & Hastie, 1992). Mock jurors who had begun to lean toward one side interpreted later evidence as more supportive of that side (Carlson & Russo, 2001). A 2022 review concluded that pretrial information, jurors’ own cognitive biases, and biased expert testimony can compound one another during a trial, and recommended that expert witnesses use sequential information-management procedures (Curley et al., 2022). The story model describes a general way of processing evidence; it cannot show how a particular jury reached its verdict.
Safeguards
Because the bias operates outside awareness, the research generally favors procedures over good intentions (Kassin et al., 2013). The principal safeguards are blind procedures, in which the person making a judgment is not told task-irrelevant information such as a suspect’s confession or another examiner’s conclusion; sequential unmasking, in which an examiner documents an analysis of the questioned evidence before seeing the reference material and receives other information only as it becomes relevant; independent verification by a second examiner unaware of the first conclusion; and documentation of what information each person received and when (Dror & Kukucka, 2021; National Research Council, 2009). Linear Sequential Unmasking–Expanded extends these principles beyond comparative disciplines by ranking information by relevance, objectivity, and biasing power, and a 2022 worksheet helps laboratories apply it in casework (Dror & Kukucka, 2021; Quigley-McBride et al., 2022).
Analogous practices exist outside the laboratory. Asking investigators to articulate why a favored theory might be wrong reduced bias in experimental studies (O’Brien, 2009). For forensic mental health experts, the review literature recommends considering alternative explanations and documenting the basis for each conclusion, while noting that the evidence for specific debiasing techniques is still limited (Neal et al., 2022). Adopting these measures implies no lack of competence or honesty; they address an influence that affects careful practitioners as well.
Limits of the Evidence
Much of this research uses students, simulated case files, or low-stakes laboratory tasks, and some influential studies are small, such as the five examiners in Dror et al. (2006). Laboratory demonstrations show that expectation can affect judgment; they do not establish how often bias changes outcomes in casework, and studies with trained examiners have sometimes found smaller or no effects (Pena et al., 2024). Effects are generally larger when evidence is ambiguous and smaller when it is clear. Several classic findings are decades old, and practices have changed since. No study can determine whether a particular investigator, interviewer, examiner, or juror was biased in a particular case, or whether a particular conclusion was wrong; at most the research identifies conditions under which the risk of bias is greater and procedures that reduce it.
Questions to Ask About This Research
When Dror et al. (2006) is cited to suggest that context shaped a forensic conclusion, counsel should ask how much evidence the finding rests on: the study involved five examiners, and Pena et al. (2024) found no effect of case information with trained fingerprint examiners. When Kukucka and Kassin (2014) is offered to show that a confession tainted a forensic comparison, counsel should ask who was studied, because the participants were laypeople rather than trained examiners. When Murrie et al. (2013) is raised against an opposing expert, counsel should ask whether it applies equally to the expert offering it, since the allegiance effect ran in both directions. When a forensic conclusion, an identification, and a confession are presented as independent corroboration, counsel should ask what each examiner, witness, or interviewer knew before reaching each result, and whether any blind or sequential procedure was documented. These questions are explained in “Reading Research Claims: A Guide for Attorneys” on this site.
Confirmation Bias and the Law
Federal and Military Standards
Expert testimony about confirmation bias is governed by Federal Rule of Evidence 702, as amended December 1, 2023, under which the proponent must show that it is more likely than not that the rule’s requirements are met, including a reliable application of reliable methods. Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993) made the judge a gatekeeper, and Kumho Tire Co. v. Carmichael (1999) extended that obligation to all expert testimony, including technical or other specialized knowledge. The 2023 amendment took effect in Military Rule of Evidence 702 in 2025 through Military Rule of Evidence 1102 (Joint Service Committee on Military Justice, 2025); practitioners should confirm current text. Military judges also apply the six factors of United States v. Houser (1993). Testimony may explain how context and expectation can affect investigations, interviews, and forensic judgments and what procedures limit that influence; it may not tell the factfinder that a particular examiner was biased or a particular witness truthful or untruthful.
Confirmation bias also bears on disclosure. The prosecution must disclose favorable evidence material to guilt or punishment (Brady v. Maryland, 1963). In Kyles v. Whitley (1995), the Supreme Court held that the prosecutor has a duty to learn of favorable evidence known to others acting on the government’s behalf, including the police, that materiality is assessed by the cumulative effect of the suppressed evidence, and that the withheld evidence in that case would have allowed the defense to attack the thoroughness and good faith of the investigation. In courts-martial, Rule for Courts-Martial 701(a)(6) requires trial counsel to disclose, as soon as practicable, known evidence that reasonably tends to negate guilt, reduce the degree of guilt or the punishment, or adversely affect the credibility of any prosecution witness or evidence.
Arizona
Arizona Rule of Evidence 702 adopted the Daubert standard effective January 1, 2012, and was amended effective January 1, 2024, to track the 2023 federal amendment (Ariz. Sup. Ct. Order R-23-0004). State v. Chapple (1983) is Arizona’s leading case on expert testimony about the reliability of memory: excluding a qualified eyewitness expert was an abuse of discretion on its facts, while admission otherwise remains discretionary (State v. Nordstrom, 2001). Experts may describe general principles but may not opine on whether a particular witness is truthful (State v. Lindsey, 1986, 149 Ariz. at 475; State v. Moran, 1986). When the State fails to preserve evidence, a defendant may request an instruction permitting the jury to infer that the lost evidence would have been unfavorable to the State (State v. Willits, 1964). The Arizona Supreme Court has described the showing required as a failure to preserve material and reasonably accessible evidence that could have tended to exonerate the accused, with resulting prejudice (State v. Glissendorf, 2014), a formulation that State v. Hernandez (2020) restated in terms of “obviously material” evidence, which it defined as evidence the State relies on, or knows the defendant will use for the defense, when the State encounters it, and has held that the instruction is not required simply because a more thorough investigation was possible (State v. Murray, 1995). Whether a particular gap in an investigation supports any instruction is a question of Arizona law for counsel and the court, not one the research answers.
Special Issues in Legal Practice
The principles described above apply with equal force to every party in a legal proceeding. 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 several propositions prosecutors can legitimately use. Confirmation bias describes a risk, not a finding that an investigation went wrong, and the effect varies with the task and procedure; one study with trained fingerprint examiners found no measurable effect of case information (Pena et al., 2024). The same research applies to defense experts and investigators, including the allegiance effects found in forensic mental health experts on both sides (Murrie et al., 2013), and an opposing expert’s opinion can be examined for what information the expert received and in what order. Prosecutors are well positioned to strengthen the evidence by encouraging blind or sequential laboratory procedures, independent verification, open-ended interviews that test more than one possibility, and records of what each investigator, interviewer, and examiner knew and when; evidence developed that way can be shown, not merely assumed, to be independent.
The research also counsels caution. Several items of evidence that appear to corroborate one another may not be independent if each was obtained with knowledge of the others, and presenting such evidence as independent confirmation can overstate its weight (Kassin et al., 2013). An investigator’s certainty, or agreement among examiners who knew the same case information, does not by itself show that the conclusion is independent or correct. Interviews conducted by someone convinced of a particular account can shape a witness’s report, and when an investigation focused early on one person, information about the leads that were not pursued may be favorable evidence subject to disclosure under Brady, Kyles, and R.C.M. 701(a)(6).
Special Issues for Defense
The research supports several propositions defense counsel can legitimately present. Expectations can influence which leads are pursued, how ambiguous evidence is rated by police officers, what questions interviewers ask, how forensic examiners and experts interpret ambiguous material, and how jurors weigh later evidence (Charman et al., 2017; Kukucka & Kassin, 2014; O’Brien, 2009; Thompson et al., 1997). Defense counsel are well positioned to develop the sequence of the investigation: when a suspect was identified, what each examiner and interviewer knew before reaching a conclusion, whether blind or sequential procedures were used, and whether apparently corroborating items were produced independently. Expert testimony explaining these processes, tied to that record, can legitimately help the factfinder weigh the evidence.
The research also counsels caution about overclaiming. Exposure to case information does not establish that a conclusion was wrong, and some studies with trained examiners have found little or no effect (Pena et al., 2024). An expert cannot tell the factfinder that a particular investigator or examiner was biased. The same research applies to the defense: defense investigators and experts form theories early, defense-retained experts have shown allegiance effects (Murrie et al., 2013), and experts generally underestimate their own susceptibility (Zapf et al., 2018). A defense expert’s own sources of information, and the order in which the expert received them, are fair subjects of cross-examination.
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
Carlson, K. A., & Russo, J. E. (2001). Biased interpretation of evidence by mock jurors. Journal of Experimental Psychology: Applied, 7(2), 91–103. https://doi.org/10.1037/1076-898X.7.2.91
Charman, S. D., Kavetski, M., & Mueller, D. H. (2017). Cognitive bias in the legal system: Police officers evaluate ambiguous evidence in a belief-consistent manner. Journal of Applied Research in Memory and Cognition, 6(2), 193–202. https://doi.org/10.1016/j.jarmac.2017.02.001
Curley, L. J., Munro, J., & Dror, I. E. (2022). Cognitive and human factors in legal layperson decision making: Sources of bias in juror decision making. Medicine, Science and the Law, 62(3), 206–215. https://doi.org/10.1177/00258024221080655
Darley, J. M., & Gross, P. H. (1983). A hypothesis-confirming bias in labeling effects. Journal of Personality and Social Psychology, 44(1), 20–33. https://doi.org/10.1037/0022-3514.44.1.20
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, Article 100161. https://doi.org/10.1016/j.fsisyn.2021.100161
Findley, K. A., & Scott, M. S. (2006). The multiple dimensions of tunnel vision in criminal cases. Wisconsin Law Review, 2006(2), 291–397.
Hill, C., Memon, A., & McGeorge, P. (2008). The role of confirmation bias in suspect interviews: A systematic evaluation. Legal and Criminological Psychology, 13(2), 357–371. https://doi.org/10.1348/135532507X238682
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., 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
Klayman, J., & Ha, Y.-W. (1987). Confirmation, disconfirmation, and information in hypothesis testing. Psychological Review, 94(2), 211–228. https://doi.org/10.1037/0033-295X.94.2.211
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
Murrie, D. C., Boccaccini, M. T., Guarnera, L. A., & Rufino, K. A. (2013). Are forensic experts biased by the side that retained them? Psychological Science, 24(10), 1889–1897. https://doi.org/10.1177/0956797613481812
National Research Council. (2009). Strengthening forensic science in the United States: A path forward. The National Academies Press. https://doi.org/10.17226/12589
Neal, T. M. S., Lienert, P., Denne, E., & Singh, J. P. (2022). A general model of cognitive bias in human judgment and systematic review specific to forensic mental health. Law and Human Behavior, 46(2), 99–120. https://doi.org/10.1037/lhb0000482
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of General Psychology, 2(2), 175–220. https://doi.org/10.1037/1089-2680.2.2.175
O’Brien, B. (2009). Prime suspect: An examination of factors that aggravate and counteract confirmation bias in criminal investigations. Psychology, Public Policy, and Law, 15(4), 315–334. https://doi.org/10.1037/a0017881
Pena, M. M., Stoiloff, S., Sparacino, M., & Schreiber Compo, N. (2024). The effects of cognitive bias, examiner expertise, and stimulus material on forensic evidence analysis. Journal of Forensic Sciences, 69(5), 1740–1757. https://doi.org/10.1111/1556-4029.15565
Pennington, N., & Hastie, R. (1992). Explaining the evidence: Tests of the story model for juror decision making. Journal of Personality and Social Psychology, 62(2), 189–206. https://doi.org/10.1037/0022-3514.62.2.189
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, Article 100216. https://doi.org/10.1016/j.fsisyn.2022.100216
Stewart, C. K., & Kukucka, J. (2025). Cognitive bias affects perception and decision-making in simulated facial recognition searches. Behavioral Sciences, 15(8), Article 1094. https://doi.org/10.3390/bs15081094
Thompson, W. C., Clarke-Stewart, K. A., & Lepore, S. J. (1997). What did the janitor do? Suggestive interviewing and the accuracy of children’s accounts. Law and Human Behavior, 21(4), 405–426. https://doi.org/10.1023/A:1024859219764
Zapf, P. A., Kukucka, J., Kassin, S. M., & Dror, I. E. (2018). Cognitive bias in forensic mental health assessment: Evaluator beliefs about its nature and scope. Psychology, Public Policy, and Law, 24(1), 1–10. https://doi.org/10.1037/law0000153
Legal Authorities
Brady v. Maryland, 373 U.S. 83 (1963).
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999).
Kyles v. Whitley, 514 U.S. 419 (1995).
State v. Chapple, 135 Ariz. 281, 660 P.2d 1208 (1983).
State v. Glissendorf, 235 Ariz. 147, 329 P.3d 1049 (2014).
State v. Hernandez, 250 Ariz. 28, 474 P.3d 1191 (2020).
State v. Lindsey, 149 Ariz. 472, 720 P.2d 73 (1986).
State v. Moran, 151 Ariz. 378, 728 P.2d 248 (1986).
State v. Murray, 184 Ariz. 9, 906 P.2d 542 (1995).
State v. Nordstrom, 200 Ariz. 229, 25 P.3d 717 (2001).
State v. Willits, 96 Ariz. 184, 393 P.2d 274 (1964).
United States v. Houser, 36 M.J. 392 (C.M.A. 1993).
Fed. R. Evid. 702.
Mil. R. Evid. 702.
Mil. R. Evid. 1102.
Rule for Courts-Martial 701(a)(6).
Ariz. R. Evid. 702.
Ariz. Sup. Ct. Order R-23-0004 (amending Ariz. R. Evid. 702, effective Jan. 1, 2024).
Related topic papers: Suggestibility of Children’s Memory · External Sources of Memory Distortion · Memory Expert Testimony · Reconstruction and Recall · Hindsight Bias · Consistency Bias
Related topics from Arizona Forensic Services: False Confessions · Child Abuse Allegations