Researchers estimate that between 1% and 5% of criminal convictions in the United States are wrong. Applied to the roughly 1.25 million people held in state and federal prisons, that range puts the number of innocent people currently incarcerated somewhere between 12,500 and 62,500. The exact figure is unknowable, because most wrongful convictions are never identified. What can be counted are the exonerations: more than 4,000 people have been officially cleared since 1989, and the number climbs every year.
Where the Estimates Come From
Measuring wrongful convictions is one of the harder problems in criminal justice research. You can count the people who have been exonerated, but that only captures cases where someone had the evidence, the legal help, and the luck to get their conviction overturned. Researchers call the rest the “dark figure” of the justice system.
A Department of Justice-funded review of the available scholarship concluded that the most widely accepted estimates fall between 1% and 5% of all convictions.1Office of Justice Programs. Estimating the Prevalence of Wrongful Conviction The range is wide because different studies use different methods and look at different categories of crime. A 2014 study in the Proceedings of the National Academy of Sciences applied survival analysis to death penalty cases and estimated that at least 4.1% of death-sentenced defendants are innocent, a figure the authors described as conservative.2National Center for Biotechnology Information. Rate of False Conviction of Criminal Defendants Who Are Sentenced to Death
Even the low end produces staggering numbers. The U.S. prison population stood at roughly 1.25 million at the end of 2023.3Bureau of Justice Statistics. Prisons Report Series: Preliminary Data Release A 1% error rate means more than 12,000 innocent people behind bars. At 5%, the figure passes 60,000. That’s the best-available math on a problem the system was not designed to measure.
How Many Have Been Exonerated
The National Registry of Exonerations, housed at the University of California, Irvine, tracks every known case of a person officially cleared of a crime after conviction. Through the end of 2024 it had recorded 3,646 individual exonerations since 1989.4The National Registry of Exonerations. 2024 Annual Report The count has since crossed 4,000 and continues to grow.5National Registry of Exonerations. National Registry of Exonerations
Together, these exonerees have lost more than 32,000 years to wrongful imprisonment.5National Registry of Exonerations. National Registry of Exonerations The overall average time served before exoneration is roughly nine years, but the trend is climbing. People exonerated in 2024 had spent an average of 13.5 years in prison before being cleared.4The National Registry of Exonerations. 2024 Annual Report The longer averages reflect legal teams taking on older, more entrenched cases that would have been considered hopeless a generation ago.
Documented exonerations are the floor of the problem, not the ceiling. They count only the people who successfully navigated a post-conviction process requiring new evidence, legal representation, and often years of litigation.
Which Crimes Produce the Most Wrongful Convictions
Wrongful convictions are not evenly distributed across offense types. Homicide and sexual assault cases dominate the exoneration records. Both categories tend to rely on eyewitness testimony, and both carry sentences high enough to attract intensive post-conviction scrutiny from defense attorneys, advocacy groups, and journalists. DNA evidence in sexual assault cases has also given innocent defendants a reliable way to prove what happened years after the original trial.
A quieter but growing category involves “no-crime” cases, particularly in drug prosecutions. These occur when someone pleads guilty to drug possession and later lab testing reveals the seized material was not an illegal substance. Because most criminal cases are resolved through plea deals rather than trials, defendants routinely accept guilty pleas to avoid the risk of a longer sentence after conviction at trial. That math makes sense even to innocent people facing mandatory minimums, and it means these errors can stay buried for years.
Misdemeanor cases almost certainly hide a large number of wrongful convictions too, but they are the least studied. The sentences are shorter, the cost of challenging a conviction often exceeds the sentence itself, and no one is funding DNA testing on a case that carried 60 days in jail. The research focus on serious felonies creates a blind spot in lower-level courts.
Leading Causes
Wrongful convictions rarely trace to a single failure. Most involve several breakdowns that compound each other. A few causes appear again and again across the case files.
Eyewitness Misidentification
Mistaken identification by a witness or victim is the single largest driver. Among DNA exonerations tracked by the Innocence Project, 69% involved eyewitness misidentification.6Innocence Project. How Eyewitness Misidentification Can Send Innocent People to Prison Human memory is far less reliable than most people assume, and the conditions under which crimes are witnessed — poor lighting, high stress, cross-racial identification — make honest mistakes common. Suggestive lineup procedures make the problem worse.
Official Misconduct
Police and prosecutorial misconduct show up in a disturbingly large share of cases. The National Registry of Exonerations reported that official misconduct was a contributing factor in 57% of all documented exonerations in its most recent reporting year.7National Registry of Exonerations. Exonerations in 2025 The category covers suppressing evidence favorable to the defense, coercing witnesses, fabricating reports, and running suggestive identification procedures.
Flawed Forensic Evidence
Forensic techniques once treated as unimpeachable have been steadily discredited. Flawed forensic evidence is the second most common contributing factor in wrongful conviction cases, appearing in nearly a quarter of all exonerations since 1989.8Innocence Project. Misapplication of Forensic Science The problem is not only bad science. It also includes forensic analysts who exaggerated their findings on the stand, mischaracterized inconclusive results, or concealed exculpatory test outcomes.
False Confessions
It strikes most people as impossible that someone would confess to a crime they did not commit, but it happens regularly. Estimates vary by dataset, but false confessions have been identified in roughly 12% to 29% of wrongful convictions studied. Long interrogations, psychological pressure, intellectual disabilities, and youth all raise the risk. People sometimes confess to end an interrogation, believing they can straighten things out later. They almost never can.
Racial Disparities
The burden of wrongful conviction falls unevenly. Research based on National Registry of Exonerations data has found that innocent Black people are roughly 19 times more likely to be wrongfully convicted of drug crimes than innocent white people, and about seven and a half times more likely to be wrongfully convicted of murder. In murder cases, exonerations of Black defendants are nearly 50% more likely to involve police misconduct than those of white defendants.
These disparities reflect the compounding effects of the causes above: racial bias in eyewitness identification, more aggressive policing and interrogation tactics in Black communities, and prosecutorial decisions that studies have found to be influenced by the defendant’s race.
Why So Few Innocent People Get Out
The gap between the estimated population of innocent prisoners and the annual exoneration count is enormous. Fewer than 200 exonerations happen each year, while thousands of innocent people likely remain incarcerated. The legal system is built around finality, and that principle works against anyone trying to prove innocence after the verdict.
To reopen a case, a convicted person generally has to present new evidence that was not available at trial. Courts do not re-weigh the old evidence or second-guess the jury. An “actual innocence” claim requires the defendant to affirmatively demonstrate innocence, which effectively flips the original burden of proof. Meeting that standard without physical evidence, surviving witnesses, or competent counsel is close to impossible for most incarcerated people.
The practical barriers are just as steep. Many states impose tight filing deadlines for post-conviction relief, sometimes only a few months after sentencing. Biological evidence that could be tested for DNA is often lost or destroyed. Public defender offices handling these cases are chronically underfunded. And for the many convictions that come from guilty pleas, the legal avenues for challenge are narrower still, because the defendant formally waived many appellate rights in the plea agreement.
Conviction integrity units, specialized teams inside prosecutors’ offices that review past cases for error or misconduct, have helped close the gap in some places. Illinois, Texas, and New York have been leaders, which is part of why those states report higher exoneration numbers. Most counties in the country still lack any formal review mechanism, so whether an innocent person has a realistic path to freedom depends heavily on geography.
Compensation After Release
Getting out is not the end of the story. Federal law caps compensation at $50,000 per year of wrongful imprisonment, or $100,000 per year for those wrongfully sentenced to death, and it applies only to federal convictions.9Office of the Law Revision Counsel. 28 USC 2513 – Unjust Conviction and Imprisonment
For state convictions, compensation depends entirely on where the wrongful conviction occurred. Thirty-eight states and the District of Columbia have wrongful conviction compensation statutes, with payments that vary widely.10National Registry of Exonerations. Compensation That leaves twelve states where exonerees have no statutory right to compensation at all. In those states, an innocent person can spend decades in prison and walk out with nothing. Some exonerees have won large civil rights settlements, but litigation takes years and is not available to everyone.