The emergence of forensic DNA analysis in the late 1980s has enhanced the criminal justice system’s ability to find the truth. In addition to facilitating the identification and conviction of the guilty, the DNA testing has also exposed a large and growing number of cases in which innocent people were convicted of crimes they did not commit. For the first time, the criminal justice system now has a body of cases in which there is scientific proof that the truth-finding mechanisms of the system failed. Jump-started by the DNA cases, recognition of wrongful convictions has expanded to include cases without any DNA as well. The study of these wrongful convictions has revealed numerous causes of errors related to the way evidence is collected and cases are tried. A commonality shared by almost all the wrongful conviction cases is the presence of a variety of cognitive distortions or biases that can lead investigators, litigators, judges, and juries astray.
Scope of the Problem of Wrongful Conviction
The American criminal justice system has historically prided itself on taking great precautions to guard against wrongly convicting the innocent. The American courts and commentators have long espoused a philosophy of caution, expressed in the maxim that it is better to let 10 (or 100) guilty people go free than to convict one innocent person. Nonetheless, there has never been real doubt that the system occasionally errs.
Even prior to the DNA revolution, scholars sought to identify wrongful convictions. In 1932, Edwin Borchard identified what he believed to be 65 wrongful convictions in serious cases. More recently, in 1987 and 1992, Hugo Bedau, Michael Radelet, and Constance Putnam identified more than 400 wrongful convictions in cases potentially subject to capital punishment. These and other similar efforts, however, were subject to challenge by skeptics, who doubted innocence in some of the cases, and even when accepted, the cases were largely dismissed as anomalies rather than symptoms of systemic flaws.
The DNA cases changed this. The DNA cases presented unassailable scientific proof of error. They also demonstrated that errors have occurred not just in those cases where proof of guilt appeared tenuous but also in cases where the evidence of guilt had appeared overwhelming. Moreover, they revealed that wrongful convictions are more prevalent than previously thought and that they reflect systemic flaws.
Determining a precise wrongful conviction rate is very difficult, as it is impossible to identify the whole body of erroneous convictions. To compound the problem, establishing an acceptable definition of “wrongful conviction” is itself difficult, especially in cases that lack dispositive DNA or other conclusive scientific evidence. Not everyone “wrongly convicted” is actually or completely innocent. Rather, some individuals who are released from their convictions because of procedural errors or inadequate evidence to prove their guilt—and who are thus legally innocent—might nonetheless be factually guilty. And though they are legally innocent, they may not be completely innocent since, for example, they may be guilty of a lesser charge. Distinguishing between these categories can be challenging.
In recent years, however, new evidence has identified a significant number of individuals who were in fact completely innocent but were nonetheless convicted. Between 1989 (the year of the first DNA exoneration in the United States) and 2006, at least 189 people who had been convicted of serious crimes in America were exonerated by postconviction DNA testing. Although significant, this number reflects just the tip of what is certainly a much larger iceberg. As important as DNA can be, it is present in only a small percentage of all criminal cases, and it is preserved and available for postconviction analysis in just a fraction of that total. Hence, the DNA exoneration reveal only a small percentage of all wrongful convictions.
Examining media and other published accounts of cases, Samuel Gross and colleagues have identified 340 cases of proven wrongful convictions in serious felonies between 1989 and 2003. In each case, the wrongful conviction was established by an official governmental act finding the person not guilty through one of several procedures: dismissal of the case by the prosecution or court in light of new evidence of innocence, an acquittal after a retrial, or a pardon based on innocence. As Gross notes, even this group is substantially underinclusive, because it relies on the happenstance that the defendant was able to discover convincing new evidence of innocence and the chance that the researchers found the case to include it in the study.
Looking at the one group of cases for which there exist fairly reliable data on proven exoneration rates—capital rape-murders—Michael Risinger has developed what might be the most empirically sound estimate of a wrongful conviction rate. Risinger calculates an error rate in capital rape-murders of approximately 3.3% to 5%.
Causes of Wrongful Convictions
Examining these wrongful convictions has revealed several recurring causes of factual error in criminal cases. The causes include, among others, eyewitness error, false confessions, unreliable jailhouse snitch or informer testimony, witness perjury, faulty forensic science, police misconduct, prosecutorial misconduct, and ineffective defense counsel.
Among these, eyewitness error is by far the most prevalent, occurring in anywhere from approximately 60% to 84% of the exoneration cases. Eyewitness error typically does not involve untruthful witnesses but rather well-meaning, honest witnesses who are simply mistaken about their memory of the perpetrator or the crime. Considerable psychological research has demonstrated the fallibility of eyewitnesses and identified factors that can contribute to eyewitness error. Eyewitness memory is susceptible to contamination and distortion by suggestive police identification procedures or post-incident information. For this reason, eyewitness evidence is sometimes analogized to trace physical evidence; as with trace evidence, the fragile nature of the evidence demands care in collecting and storing the evidence. In eyewitness identification cases, care must be taken to minimize suggestiveness and contamination of eyewitness memories.
False confessions are also prominent among the causes of wrongful convictions. Although it is counterintuitive to imagine that an innocent person would confess to a crime he or she did not commit, the wrongful conviction cases demonstrate that false confessions are present in up to nearly one quarter of exoneration cases. High pressure, confrontational police interrogation tactics, such as those included in the Reid Technique of interrogation (which is taught in some form in most police jurisdictions in the United States), which is believed to be effective at eliciting confessions from the guilty, also can induce innocent people to confess. Social science research suggests that under such interrogation tactics, false confessions can be the product of rational choices.
Jailhouse informer or snitch testimony typically involves testimony offered by an individual who was, or claims to have been, incarcerated with the defendant and who claims that the defendant confessed or made incriminating statements to him or her while they were incarcerated together. Courts have long recognized that such witnesses are very unreliable, both because their criminal background suggests that they might have little regard for the truth and because they have an incentive to fabricate. Jailhouse informers often are motivated by explicit or implicit promises, or even unilateral hopes, of leniency or benefits from the prosecution in their own criminal cases if they provide useful information against another.
Numerous wrongful convictions have also rested, at least in part, on fraudulent or mistaken forensic science. Occasionally, these errors are the product of deliberate fraud. A number of high-profile instances of such fraud have been reported in recent years, in which crime laboratory analysts reported incriminating scientific test results when in fact the analysts either obtained no results, obtained nonincriminating results, or did not even run the tests (a type of fraud referred to as “dry labbing”).
More typically, however, laboratory analysts have made honest errors. In some instances, the purported forensic science has itself been fundamentally unreliable and nonscientific. To take one example, microscopic hair comparison, which was a staple of criminal prosecutions for many years, has very little scientific foundation and has now been exposed by DNA analysis as frequently incorrect. Accordingly, many crime laboratories no longer conduct microscopic hair examination.
In other instances, laboratory analysts may be influenced by expectation effects. Research has shown that when laboratory analysts are informed of the results that are expected or about other evidence in a case, this non-domain information can influence the analysts’ interpretation of ambiguous data. When told that other evidence includes or excludes a suspect, for example, analysts are more likely than otherwise to conclude that their scientific analyses are consistent with that other information.
Police and prosecutorial misconduct involves overreaching in a variety of contexts. The most common type of prosecutorial misconduct involves failure to comply with the constitutional mandate that prosecutors must disclose to the defense all material exculpatory evidence in their possession. In part, a prosecutor’s failure to comply with this mandate reflects the very difficult demands that the adversary system imposes on prosecutors. Since a prosecutor’s responsibility to convict the accused naturally encourages him or her to view the evidence in an inculpatory light, it is too much to expect that the same prosecutor would simultaneously view the evidence from the defendant’s perspective and recognize its exculpatory value.
Finally, inadequate defense counsel is a frequent cause of wrongful convictions. Indigent legal services are chronically underfunded, and the result frequently is inadequate defense investigation and a lackluster challenge to the state’s case at trial. When the defense is inadequate, the adversarial system fails to function as designed to weed out erroneous charges or protect the innocent.
These individual causes of wrongful convictions often work in conjunction with one another to produce a faulty assessment of guilt. A mistaken eyewitness identification, for example, can convince the police of a suspect’s guilt. Once convinced of guilt, the police may then set out to develop the evidence needed to obtain a conviction. They might aggressively interrogate the suspect to obtain a confession, producing incriminating statements from the suspect, leading the police and prosecutors to interpret innocent responses in an inculpatory manner, or even inducing the suspect to confess falsely. The police and prosecutors might also seek an unreliable jailhouse informant to bolster their case. Laboratory analysts, informed of the state’s theory of guilt, might interpret ambiguous data to support that conclusion. Or the police and prosecutors might otherwise cut corners or bend the rules, in the belief that doing what it takes to convict a guilty person serves the interests of justice. The result of this process is that initial assessments of guilt are reinforced, and the confidence of eyewitnesses, the police, prosecutors, and ultimately courts is bolstered in their judgments about the defendant’s guilt.
Cognitive Distortions and Biases in Wrongful Conviction
Regardless of the particular errors in a given case, a commonality in most wrongful convictions is the effect of several cognitive distortions or biases in producing a kind of “tunnel vision” that impedes accurate assessment of the facts. The most prominent is confirmation bias—the natural human tendency to seek, interpret, and recall information in ways that support existing expectations, beliefs, or hypotheses. Numerous studies have shown that when testing a hypothesis, people tend to seek information that confirms the hypothesis. In studies, people demonstrate a preference for evidence that will confirm their hypotheses over evidence that will disconfirm them, even though the latter is frequently more probative. By seeking only information that is consistent with their hypotheses, people fail to discover evidence that might disprove their hypotheses and reveal that their confirming evidence was merely coincidental. In a criminal case, this means that investigators tend to look for evidence that is consistent with their theory of guilt but tend not to look for disconfirming evidence—that is, evidence that would exonerate a suspect.
People similarly have a natural tendency to recall and interpret information in a manner that confirms their beliefs. Research shows a general tendency to overweight positive or confirmatory evidence and underweight negative or disconfirmatory evidence. In criminal cases, this tendency means that investigators and prosecutors are likely to ignore or minimize dis-confirming evidence—deeming the evidence irrelevant or the witness unreliable—while overrelying on confirming evidence—interpreting ambiguous data as exculpatory and judging incriminating witnesses and information as highly relevant and reliable.
Compounding these tendencies are phenomena such as belief persistence, also known as belief perseverance. Research shows that people are naturally disinclined to relinquish initial conclusions or beliefs, even when the bases for those initial beliefs are undermined. For example, once convinced of guilt in part because of an initial assessment that crime scene hairs bore microscopic physical characteristics that “matched” a suspect’s, investigators or prosecutors in numerous cases have persisted in their belief of guilt even after new DNA testing has proven conclusively that the hairs did not come from the defendant.
Such tunnel vision is also reinforced by other cognitive biases, such as hindsight bias, or the “knew-it-all-along effect.” Hindsight bias refers to the tendency that people have to use information obtained after an event to conclude that the eventual outcome was inevitable or more predictable than it actually was. With knowledge of an outcome, people’s memories tend to elaborate or emphasize evidence that was consistent with the outcome and minimize or discount evidence that was inconsistent.
In criminal cases, once the police, prosecutors, and courts conclude that an individual is guilty, hindsight bias would suggest that the suspect was an obvious and inevitable suspect from the beginning. In hindsight, evidence against that individual is enhanced. That hindsight assessment in turn reinforces the commitment to focus on that person as the culprit.
Similarly, hindsight bias can affect a witness’s assessment of or confidence in his or her identification of a suspect. For example, if an eyewitness had a fleeting glimpse of a perpetrator, that witness likely had a poor image or memory of the perpetrator. But if the witness subsequently viewed clear images of the suspect in a photo spread or live-person lineup and attempted an identification, the witness might replace the poor memory of the perpetrator from the crime with the clear image of the suspect from the photo spread or lineup. Although the identification might be wrong (given that the witness actually had a poor view and memory of the suspect), the witness might in hindsight draw on the clear image from the photo spread or lineup to conclude confidently that he or she had a good view and memory of the suspect and made an accurate identification, especially if the witness received any confirming feedback after making the identification.
Hindsight bias can also have a profound impact on judges called on to review the validity of convictions in post-conviction proceedings or on appeal. Typically, such courts are required to assess whether any errors committed by the prosecutor, the court, or defense counsel might have made a difference in the outcome of the case. Hindsight bias makes it naturally difficult for courts to imagine that any error might have affected the outcome. The outcome of the case—the defendant was found guilty beyond a reasonable doubt—tends to appear, in hindsight, to have been inevitable and appropriate. Under these circumstances, it is difficult for courts to conclude that any but the most egregious errors might have made a difference. This, in part, helps explain why courts can be extremely reluctant to reverse convictions, even in the face of strong evidence of error.
Wrongful Conviction Reforms
The increasing awareness in the criminal justice system of the problem of wrongful convictions has also led to an increasing interest in reforms to reduce the rate of such errors. Policymakers are interested in reforms to prevent wrongful convictions, not just because each such case is an injustice to the wrongly convicted but also because every time an innocent person is wrongly convicted, the true perpetrator eludes prosecution. Both basic fairness and public safety demand reliability in the criminal justice system. A variety of official commissions and policy-making bodies have been created in a number of jurisdictions to examine the wrongful conviction cases and develop recommendations for reforms to minimize such errors.
To date, the most progress in implementing reforms designed to minimize wrongful convictions has been made in the areas of eyewitness error and false confessions. In particular, extensive psychological research has produced a well-developed series of recommendations for improving eyewitness identification procedures. A number of law enforcement agencies throughout the country are now implementing some or all these recommendations.
Some of the more significant eyewitness identification reforms include ensuring that witnesses are properly instructed that the perpetrator might not be present in any given lineup or photo array, so that the witness does not feel compelled to pick someone in every case; properly selecting lineup or photo array “fillers” (non-suspects) so that the suspect does not stand out; presenting no more than one suspect in any given lineup or photo array; conducting the identification procedure in a “double-blind” manner—meaning that neither the witness nor the detective administering the procedure knows which individual is the suspect—so that the detective cannot even inadvertently cue the witness as to which individual to pick; and presenting photographs of lineup members sequentially, rather than simultaneously, so that the witness must rely on absolute judgments drawn from the witness’s memory rather than relative judgments based on comparing one lineup member or photograph with the others.
The most significant reform designed to prevent false confessions is a requirement that all custodial interrogations be electronically recorded from start to finish. Recording serves several purposes. It deters the police from engaging in improper coercive tactics that can produce false confessions. It also creates a clear record of what was said and done, so that lawyers, judges, and juries can more fully and accurately consider the reliability of any statements elicited during an interrogation and indeed help fact finders determine accurately what the suspect said, in his or her own words, without interpretation or paraphrasing by police witnesses. Electronic recording also protects the police from spurious claims of misconduct in the interrogation room and produces powerful evidence to help convict the guilty when a suspect freely and convincingly confesses or incriminates himself in a recording that can be played for the jury.
- Bedau, H. A., & Radelet, M. L. (1987). Miscarriages of justice in potentially capital cases. Stanford Law Review, 40, 21-179.
- Borchard, E. M. (1932). Convicting the innocent: Errors of criminal justice. New Haven, CT: Yale University Press.
- Findley, K. A., & Scott, M. S. (2006). The multiple dimensions of tunnel vision in criminal cases. Wisconsin Law Review, 2, 291-398.
- Givelber, D. (1997). Meaningless acquittals, meaningful convictions: Do we reliably acquit the innocent? Rutgers Law Review, 49, 1317-1396.
- Gross, S. R., Jacoby, K., Matheson, D. J., Montgomery, N., & Patil, S. (2005). Exonerations in the United States, 1989 through 2003. Journal of Criminal Law and Criminology, 95, 523-560.
- Radelet, M. L., Bedau, H. A., & Putnam, C. E. (1992). In spite of innocence: Erroneous convictions in capital cases. Boston: Northeastern University.
- Risinger, D. M. (in press). Innocents convicted: An empirically justified factual wrongful conviction rate. Journal of Criminal Law and Criminology, 97.
- Rosen, R. A. (2006). Reflections on innocence. Wisconsin Law Review, 2, 237-290.
- Saks, M. J., & Koehler, J. J. (2005). The coming paradigm shift in forensic identification science. Science, 309, 892-895.
- Scheck, B., Neufeld, P., & Dwyer, J. (2000). Actual innocence: Five days to execution and other dispatches from the wrongly convicted. New York: Doubleday.