Some forensic evidence is highly reliable and some is not. DNA from a single source is the strongest method there is. Fingerprints are valid, with known error rates. Bitemark matching, hair analysis and some firearms testimony have weak support in science. Misused forensic science played a part in 43% of the first 375 U.S. DNA exonerations. That is why courts and labs are tightening how this evidence is used.
TV shows make forensic science look certain. The research tells a more mixed story. Lawyers, jurors, lawmakers and business leaders all rely on what investigations find. For them, knowing the gap between strong and weak methods matters.
The 2009 Report That Changed the Conversation
In 2009, the National Academy of Sciences published a major review of forensic science in the United States. It found that only nuclear DNA analysis had been shown to reliably connect evidence to a single person with a high degree of certainty. Many other fields had never been put through strict scientific tests.
The report also found a broken-up system. Most labs and workers did not have to be accredited or certified. The authors called for several fixes. They wanted a national forensic science institute that stood on its own. They wanted labs that work apart from police and prosecutors. They also asked for required accreditation and for testimony that states the limits of each method.
How Each Method Stacks Up
In 2016, the President’s Council of Advisors on Science and Technology took the next step. It asked whether each method had been tested in ways that show how often examiners get it wrong. The PCAST report reached very different answers for different fields.
- Single-source and simple DNA mixtures. Valid and reliable in science.
- Complex DNA mixtures. Reading them by human judgment was not valid. Software methods held up only in narrow cases.
- Latent fingerprints. Valid, but with a real rate of false matches. It could be as high as 1 in 306 in an FBI study. It could be as high as 1 in 18 in a study by another crime lab.
- Firearms identification. Not yet proven at the time. The one fitting study showed a false match rate of about 1 in 66.
- Bitemarks. Far from valid. False matches were often above 10%.
- Hair comparison. Not shown to be valid. In an FBI review, hairs judged to match came from different people in about 11% of cases (9 of 80).
The Human Cost of Weak Evidence
The Innocence Project tracked 375 DNA-based exonerations in the United States from 1989 to 2020. Misapplied forensic science played a role in 43% of them. Only eyewitness mistakes played a role more often.
The problem is not just in the past. The National Registry of Exonerations reported 147 exonerations in 2024. Of those, 29% involved false or misleading forensic evidence. On average, the people cleared that year had lost 13.5 years in prison.
One review shows how far errors can spread. In 2015, the FBI admitted that its examiners gave erroneous statements in 257 of 268 trial cases that used hair analysis under a microscope. The errors showed up in 33 of 35 death penalty cases reviewed.
How Courts Are Responding
Judges now look hard at how experts word their findings, most of all in gun cases. In 2023, Maryland’s highest court ruled in Abruquah v. State that an examiner could not testify without qualification that a bullet came from a specific gun. The examiner could still say the marks were consistent or not consistent.
Other courts have taken like steps. An Oregon appeals court and an Illinois trial court both limited firearms testimony in 2025. In June 2026, Maryland’s Supreme Court made clear that Abruquah applied to that case alone. It was not a blanket ban. That shows the law here is still settling.
Bitemark evidence has faced the strongest pushback. Texas’s Forensic Science Commission called for a pause on its use in 2016. That came after a man was cleared after 28 years in prison. A 2022 NIST draft review found that bitemark analysis lacks a sufficient scientific foundation.
The Push for Standards and Accreditation
Real progress has followed the criticism. The Organization of Scientific Area Committees, run by NIST, now lists 264 standards on its registry. More than 300 labs and agencies have signed up to use them. The Bureau of Justice Statistics found that 88% of publicly funded crime labs were accredited in 2014. That was up from 70% in 2002.
Gaps remain. In that same BJS survey, almost every lab ran skill tests. But only about 10% used blind tests, where analysts do not know they are being tested. Blind tests give a better read on day-to-day accuracy.
Reformers inside the field have pushed hardest for these changes. Business Insiders has profiled a reformer with a plan to fix the forensic quality problem. He is part of a growing group of experts who say quality systems, not just new tools, decide whether evidence can be trusted. More legal coverage is in the law section.
What This Means Outside the Courtroom
The same lessons apply to company probes, insurance claims and internal audits. Before relying on any analysis, decision makers should ask three questions.
- Has this method been tested in studies that measure how often it is wrong?
- Was the lab or examiner accredited and skill tested?
- Does the finding state its limits, or does it claim a certainty the science cannot back up?
Frequently Asked Questions
What is the most reliable forensic evidence?
DNA from a single source or a simple mixture is the most reliable. Federal science reviews found it can link evidence to one person with a high degree of certainty.
How often is forensic evidence wrong?
It depends on the method. Fingerprint false match rates in two key studies could be as high as 1 in 306 and 1 in 18. Bitemark matches had false match rates above 10%.
Is firearms evidence still allowed in court?
Yes, but courts more and more limit how examiners word it. Several now bar experts from claiming for certain that a bullet came from one specific gun.
The Bottom Line
Forensic science is not one thing. It is a set of methods with very different track records. DNA stands on solid ground. Pattern-matching fields are getting better under pressure from scientists, courts and reformers. Anyone who relies on forensic findings should ask how well the method has been tested, not just what the expert concluded.
