A lead is not evidence of guilt
Investigations generate many leads: a tip line call, a database hit, a vehicle sighting, an inconsistency in a statement. Most are eliminated. When coverage reports that police are "looking at" someone, it usually describes a lead at an early stage of testing. Our guides treat every lead as provisional until it has passed through the checks described below and, ultimately, through a court.
The scientific reviews that reset the field
Two U.S. reports shape how we describe forensic evidence. In 2009 the National Research Council of the National Academy of Sciences published Strengthening Forensic Science in the United States: A Path Forward. It found that, apart from nuclear DNA analysis, many forensic disciplines had not been rigorously shown to consistently connect evidence to a specific individual or source. In 2016 the President's Council of Advisors on Science and Technology (PCAST) assessed "feature-comparison" methods such as bite marks, firearms, footwear and latent prints, and called for empirical studies of foundational validity and error rates.
In response, the National Institute of Standards and Technology (NIST) administers the Organization of Scientific Area Committees for Forensic Science (OSAC), which maintains a registry of forensic standards. Laboratories commonly seek accreditation to ISO/IEC 17025, the international standard for testing and calibration laboratories.
Methods and their documented limits
| Method | What it can support | Documented limit |
|---|---|---|
| Single-source DNA profile | Strong association with a source, expressed statistically | Does not show how or when DNA was deposited. |
| Mixed DNA samples | Possible inclusion or exclusion | Interpretation of complex mixtures was flagged by PCAST as needing validation. |
| Latent fingerprints | Association by examiner comparison | Misidentification occurs; see the Mayfield case. |
| Forensic genetic genealogy | Investigative lead toward a family line | Not an identification; requires confirmatory testing. |
| Eyewitness identification | Recollection of events and people | Memory is malleable; procedure strongly affects accuracy. |
| Bite-mark comparison | Very limited | PCAST found it lacked foundational validity. |
Case note: Brandon Mayfield, 2004
Wrongly identified, released After the March 2004 Madrid train bombings, FBI examiners linked a latent print recovered in Spain to Brandon Mayfield, a lawyer in Portland, Oregon. He was detained as a material witness in May 2004. Spanish National Police later identified the print as belonging to an Algerian national, and Mayfield was released. The U.S. Department of Justice Office of the Inspector General's 2006 report attributed the error partly to the unusual similarity of the prints and to examiners' reasoning backward from features of Mayfield's print. The U.S. government settled his lawsuit and issued an apology.
Case note: Sally Clark, England, 1999–2003
Conviction quashed Sally Clark was convicted in 1999 of murdering her two infant sons. At trial, paediatrician Roy Meadow gave a figure of "1 in 73 million" for two sudden infant deaths in one family, a figure that assumed the deaths were independent events. The Royal Statistical Society publicly criticised the misuse of statistics. Her conviction was quashed by the Court of Appeal in 2003, chiefly because microbiological test results had not been disclosed to the defence. The case is a lasting caution about how probabilities are presented to juries.
Forensic genetic genealogy
Genetic genealogy compares crime-scene DNA with profiles in consumer databases whose users have consented to law enforcement matching. It produces a family tree, not a name. In 2019 the U.S. Department of Justice issued an interim policy limiting its use mainly to violent crimes and requiring confirmatory direct DNA comparison before charges. When we report a genealogy-assisted arrest, we note that the arrest follows a confirmatory test, and that the charge still has to be proven.
Eyewitness procedure
The National Academy of Sciences' 2014 report Identifying the Culprit recommended double-blind lineups, standardised instructions and recording a witness's confidence at the time of the first identification. Mistaken eyewitness identification is a contributing factor in a large share of DNA-based exonerations documented by the Innocence Project and the National Registry of Exonerations.
Terminology notes
"Match"
Many laboratories now avoid this word. Reports may instead give a likelihood ratio or state that a person "cannot be excluded". We reproduce the laboratory's wording.
"Person of interest"
An informal term with no legal status. It does not mean a person is suspected of a crime in any formal sense.
"Cold hit"
A database association made without a prior suspect. It is a lead that must be corroborated.
Methods do not solve cases on their own. People interpret them, and the best protection against error is transparency about that interpretation.