From Pervasive Beliefs To Wrongful Convictions

Popular series like “Making a Murderer” and “The Confession Tapes” have raised public awareness of the possibility of wrongful convictions. Can cognitive psychology resolve why evidence of innocence is overlooked in cases of wrongful convictions?

The many exonerations exposed by initiatives such as the Innocence Project [1] suggest that the problem of wrongful convictions is much more widespread than a few notable examples. The national registry of exonerations lists 2363 exonerations in the US, and wrongful convictions continue to be uncovered in other countries as well [2]. The questions of why wrongful convictions occur, and how they could be prevented, are the focus of this article.

When wrongful convictions come to light, the focus tends to be on the evidence which provided false support for the defendant’s guilt. The Innocence Project website lists several common sources of such evidence, including false confessions, misapplication of forensic science, or eyewitness misidentification [1]. Many times, with wrongful convictions, there is also evidence that could have proven the defendant’s innocence. One such case is the Schiedam Park murder in the Netherlands, where a ten-year old girl was murdered and her friend was assaulted but survived [3]. The suspect became the focus of the investigation after a police employee had used the police database to investigate a private matter of sexual advances towards a minor. The man he was investigating had called the emergency services on the day of the murder after the surviving friend had asked him for help. He had therefore been heard as a witness, but following the new incident with a minor, he became a suspect of the attack. The surviving victim of the attack had described the perpetrator, but that description did not match the suspect’s appearance. In addition, several witnesses described having seen another man in the park, matching the description given by the surviving victim. DNA that could not be the suspect’s was found at the crime scene. The existence of that DNA was denied by the public prosecutor to the defence and the judges [4]. Although the suspect initially denied having attacked the children, he later confessed. However, the suspect’s description of the crime was very different from what the surviving victim had described. The suspect soon retracted his confession, but nevertheless continued to be intensively interrogated over the next four months, and was ultimately convicted [3]. As the actual perpetrator later confessed the crime, and his DNA matched the DNA found at the crime scene, the original conviction was recognised as wrongful. In response, an evaluation of the Dutch public prosecution was conducted. It concluded that tunnel vision had played a role in the wrongful conviction; evidence supporting the suspect’s innocence had been disregarded or misinterpreted [5].

Faulty Reasoning

To answer the question of why exonerating evidence is sometimes overlooked, we can start by looking at how logical reasoning should ideally work, and how it can go wrong. In his pivotal work on scientific reasoning, Popper [6] stressed the need to attempt falsification; trying to disprove one’s theory or hypothesis. According to Popper, finding evidence that confirms your theory only goes so far. Even if you have 100 pieces of evidence that support your theory, it would only take one contradicting observation to prove your theory cannot be true. Popper therefore argued that, in addition to looking for evidence that confirms your theory, it is also important to look for evidence that could refute the theory. If several serious attempts at disproving your theory have failed, it is more likely to be correct.

Unfortunately, it seems that focussing on disproving information is easier said than done. People have a natural tendency to try to confirm their current theory, and so often disproving evidence is disregarded or not considered. One demonstration can be seen in the Wason selection task [7], where people are given a conditional rule and four cards. For instance, the rule might be “if one side of the card has an even number, then the other side must be red”, and the four cards shown could be 2, 5, red, and blue. Participants are then asked to select those cards, and only those cards, which need to be turned around in order to determine whether the rule is true or false. The majority of participants in the original experiment chose the equivalent of the cards 2 and red, or only 2. The correct answer is 2 and blue. If the reverse side of the 2 card is not red, then the rule is false. The same conclusion holds if the reverse side of the blue card is an even number. The fact that people rarely provide the correct answer demonstrates the human tendency to favour confirming evidence. That tendency is contrary to the principles of logical thinking. The Wason card selection test has previously been used to demonstrate the difficulty of applying logical thinking to criminal fact-finding [8].

A common interpretation of Wason’s classical demonstration is to view it as an example of confirmation bias, the tendency to seek and interpret evidence in line with an existing belief, thereby paying disproportionately less attention to evidence that contradicts that belief. As concluded by Nickerson [9], confirmation bias is deserving of attention due to its abundant and pervasive presence in several contexts. The presence of confirmation bias may be difficult to determine, but how contradicting evidence is used may provide an indication. For instance, in the Schiedam Park murder, the police believed the suspect was guilty. The fact that he initially came to their attention after soliciting a minor for sexual acts likely contributed to that belief. Despite the discrepancy between the suspect and the description given by the victim, that belief was maintained. When a belief is maintained in the face of contradicting evidence, it is known as belief perseverance, a process thought to be a precursor to confirmation bias. The initial belief, in this case, in the suspect’s guilt, could thus be maintained. Furthermore, the suspect’s confession differed from how the victim had described the attack, but did not cause the police to change their belief in the suspect’s guilt. Even after the confession was retracted, it continued to be seen as evidence of the suspect’s guilt. The fact that the confession was given more weight than the recanting of the confession can also be interpreted as an indication of confirmation bias, as incriminating and exculpatory evidence was weighted differently. Similarly, the continued interrogations of the suspect over a period of four months after he retracted his confession can also be seen as a disproportionate focus on finding incriminating evidence. The Schiedam Park murder therefore serves as an example of how confirmation bias can affect the interpretation of evidence in light of an existing belief. That such a belief can be pervasive is supported by the fact that some prosecutors involved in proven wrongful convictions maintained a belief in the suspect’s guilt [10]. It has also come to our attention that, in the Schiedam murder case, some of the police officers involved still believe the initial conviction was correct.

Confirmation bias in legal decision-making

Confirmation bias promotes adherence to the original belief by favouring information that supports it. Contradictory evidence may be interpreted so that it is seen as supporting the initial belief, or may be weighed differently. Confirmation bias has been researched in various contexts relevant to miscarriages of justice. For instance, Kassin, Goldstein and Savitsky [11] found that mock interrogators who expected a mock suspect to be guilty used more guilt-presumptive questions, and exerted more pressure, than those who expected the mock suspect to be innocent. Despite the equal number of guilty suspects in each of the conditions, neutral observers, who were shown the recorded interrogations, also perceived the mock suspects in the guilty expectation condition as more guilty. Thus, the belief of the interrogators influenced their behaviour, as well as the subsequent perception of the suspect. Dror, Charlton & Peron [12] gave fingerprint examiners prints they had previously concluded to be a match. The examiners now also received context information that suggested the fingerprints they were examining were not a match. The majority of the experts changed their mind about the fingerprints matching. Ask and Granhag [13] presented both police officers and students with a short case about a murder, and presented either a motive for the prime suspect or an alternative scenario. The authors found that initial beliefs about guilt influenced the subsequent interpretation of evidence, and that police investigators appeared to be less open to the suggestion of an alternative scenario than students. In short, the effects of prior beliefs on subsequent reasoning has been amply demonstrated in relation to criminal investigations.

Rassin and colleagues [5] hypothesised that criminal proceedings reinforce confirmation bias. They tested their hypothesis by presenting law students with a case and asking them to determine the suspect’s guilt, after which they could choose from a list of further investigative measures. Half of the suggested measures were aimed at gathering further evidence against the suspect. The other half were aimed at obtaining exonerating information. Participants did not favour incriminating evidence, but rather favoured those investigations which supported their initial determination of guilt. Thus, participants seemed to show confirmation bias rather than a guilt bias.

The influence of existing beliefs on legal decisions has also been investigated among judges [14]. One group of judges received information that was mainly incriminating prior to the start of a trial. Another group of judges received pre-trial information which was less incriminating. Both groups then read identical records of the trial proceedings. When asked to deliver a verdict, 82% of judges in the first group thought the accused should be found guilty, compared to only 53% in the second group. It therefore seems judges are also prone to belief perseverance.

Based on these findings, it would seem that the information initially presented to judges, and the extent to which it incriminates the suspect, could influence the outcome of a case. For instance, in the Netherlands, the judge receives the case file of the police investigation, which is likely to be biased towards the guilt of the suspect, as only cases with sufficient evidence against the suspect will be brought before the judge [15]. Although the defence lawyer can make requests for investigative measures, his ability to argue for the innocence of the suspect remains limited during the pretrial proceedings. The judge reads the case file in preparation for the trial, and is then expected to remain impartial in the decision-making process. Based on the biases described above, that seems to be an extremely difficult task.


Potential remedies

One question that then arises is what can be done in order to limit the influence of an initial belief on later decisions. A method which has previously been researched is the use of alternative scenarios. O’Brien [16] found that participants who formulated a hypothesis about a favoured suspect showed more confirmation bias than those who did not. Although merely considering alternative scenarios did not counter confirmation bias, when participants were asked to think about why their hypothesis might be wrong, participants showed less confirmation bias in their selection and interpretation of evidence. Active consideration of evidence refuting a hypothesis may be necessary to safeguard against confirmation bias.

Rassin [17] investigated the use of a pen-and-paper tool to consider alternative scenarios in the legal decision-making process. Participants were asked to read a case vignette, which was either focused on one suspect, or which also introduced an alternative scenario (the vignettes were based on the above study by Ask and Granhag [13]). Some participants who were shown the alternative scenario version of the case were also given a pen-and-paper tool to evaluate the evidence, and asked to imagine someone else committed the crime. Those who made use of the pen-and-paper tool had a significantly lower rating of the suspect’s guilt than the other two conditions. The author argued that the act of considering how well each piece of information fits with an alternative scenario may raise awareness of the fact that multiple explanations of the existing evidence are possible, ultimately lowering the conviction of guilt. When the study was conducted using criminal justice professionals, Rassin [17] also found that using the pen-and-paper tool led to a more balanced consideration of the evidence in light of the alternative scenarios. Hence, the alternative scenario needs to be actively assessed, rather than merely considered, in order to be beneficial. When relating these findings to the Schiedam Park murder, it can be argued that although the evidence contained indications for an alternative scenario, such as the DNA found and the lack of guilty knowledge within the confession, the scenarios were not actively considered during the course of the investigation.

In the studies discussed above several researchers have demonstrated that the mere consideration of an alternative scenario is not sufficient to prevent confirmation bias. However, it appears from the findings by O’Brien [15] that thinking specifically about why a hypothesis might be false does reduce the influence of the initial hypothesis. That brings us back to Popper’s advocacy of falsification, and arguments made for the application of logical reasoning to legal fact-finding [6,8]. Although it is possible that one scenario is falsified through the confirmation of another (e.g., when DNA proved the guilt of the actual perpetrator in the Schiedam murder case, it disproved the guilt of the initial suspect), that is not always the case. For instance, evidence that confirms one scenario might not disprove an alternative scenario. Thus, perhaps the focus should be on trying to falsify scenarios rather than trying to see which scenario has the most support. The scenario with the fewest successful attempts at falsification would then be most likely. The question then becomes how, taking into consideration confirmation bias, one can be encouraged or trained to attempt falsification, and look for evidence that could disprove one’s initial theory. Shaw and colleagues [18] recently presented some promising results: confirmation bias was considerably reduced in their participants who completed an educational computer game about biases. Nevertheless, more research on ways to counter confirmation bias, specifically in the context of legal decision-making, is needed.


When considering confirmation bias as a contributing cause of miscarriages of justice, the need for a way in which its influence can be reduced becomes apparent. Although several researchers have looked at the effects of confirmation bias in various areas, and the body of research on how it can be prevented is continually growing, it appears no definitive solution has been offered. Therefore, further research into possible countermeasures against confirmation bias should be welcomed. One promising approach to be further investigated is encouraging legal decision-makers to focus on falsifying rather than confirming their hypotheses.


  2. For the national registry of examinations, see For examples from the Netherlands, see For examples from the UK, see
  3. Van Koppen, P.J. (2003). De Schiedammer parkmoord: Een rechtspsychologische reconstructie. [The Schiedam Park murder: A legal psychological reconstruction] Nijmegen: Ars Aequi
  4. Haan, B. (2006) Justitie hield crucial bewijs achter in Schiedammer Parkmoord.[Prosecution withheld crucial eviden in the Schiedam Park murder]Jaarboek Onderzoeksjournalistiek. Retrieved from:
  5. 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, 231-246. Doi: 10.1002/jip.126
  6. Popper, K. (1959). The logic of scientific discovery. New York: Routledge
  7. Wason, P.C. (1968). Reasoning about a rule. Quarterly Journal of Experimental Psychology, 20, 273-281. Doi: 1080/14640746808400161
  8. Rassin, E. (2018). Fundamental failure to think logically about scientific questions: An illustration of tunnel vision with the application of Wason's Card Selection Test to criminal evidence. Applied Cognitive Psychology32, 506-511. Doi: 10.1002/acp.3417
  9. Nickerson, R.S. (1998) Confirmation bias: A ubiquitous phenomenon in many guises. Review of general psychology, 2, 175-220. Doi: 10.103/2F1089-2680.2.2.175
  10. Burke, A. (2007). Neutralizing cognitive bias: An invitation to prosecutors. New York University Journal of Law Liberty, 2, 512-530. Retrieved from:
  11. Kassin, S. M., Goldstein, C. C., & Savitsky, K. (2003). Behavioral confirmation in the interrogation room: On the dangers of presuming guilt. Law and human behavior27, 187-203. Doi: 1023/A:1022599230598
  12. Dror, I. E., Charlton, D., & Péron, A. E. (2006). Contextual information renders experts vulnerable to making erroneous identifications. Forensic science international156, 74-78. Doi: 10.1016/j.forsciint.2005.10.017
  13. Ask, K., & Granhag, P. A. (2005). Motivational sources of confirmation bias in criminal investigations: The need for cognitive closure. Journal of Investigative Psychology and Offender Profiling2, 43-63. Doi: 10.1002/jip.19
  14. Schünemann (1983), as cited in, Schünemann, B. & Bandilla, W. (1989) Perseverance in courtroom decisions. In H. Wegener, F. Lösel, & J. Haisch (Eds.) Criminal behavior and the justice system: Psychological perspectives (pp. 181 – 192). New York: Springer
  15. Crombag, H.F.M. (2017) Invloeden op rechterlijke beslissingen.[Influences on judicial decisions] In P.J. van Koppen, J.W. de Keijser, R. Horselenberg, & M. Jelicic (Eds.) Routes van het recht [Routes of law], (pp. 385-397). Den Haag: Boom Juridisch.
  16. 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, 315-334. Doi: 10.1037/a0017881
  17. Rassin, E. (2018). Reducing tunnel vision with a pen‐and‐paper tool for the weighting of criminal evidence. Journal of Investigative Psychology and Offender Profiling, 1, 1-7. Doi: 10.1002/jip.1504
  18. Shaw, A., Kenski, K., Stromer-Galley, J., Mikeal Martey, R., Clegg, B. A., Lewis, J. E., Folkestad, J.E., Strzalkowski, T. (2018). Serious efforts at bias reduction: The effects of digital games and avatar customization on three cognitive biases. Journal of Media Psychology: Theories, Methods, and Applications, 30, 16-28. Doi: 10.1027/1864-1105/a000174

article author(s)