Explainers

Reading the Crime Scene: What Forensics Can Reveal — and What It Can't

A crime scene is a record. Every object that was touched, every drop that fell, every mark left behind is a sentence in an account of what happened — if you know how to read it. Forensic science is the discipline of reading that account carefully, and its first principle is more than a century old: every contact leaves a trace. But reading a scene honestly means knowing two things at once — what the evidence can genuinely reveal, and where confident-sounding “scene reading” has gone catastrophically wrong and helped convict the innocent. Both halves matter, and we hold to both.

The founding idea: every contact leaves a trace

In the early 1900s the French criminalist Edmond Locard articulated what’s now called Locard’s exchange principle: when two things come into contact, each takes something from the other and leaves something behind. A burglar carries away carpet fibers and leaves behind skin cells; a struggle transfers hair, blood, and trace material in both directions. The principle is the reason a scene is worth reading at all — it promises that an encounter leaves a physical signature, even when no one saw it happen. Reconstruction is the work of recovering that signature and asking what sequence of events could have produced it.

What a scene can genuinely tell you

Good reconstruction doesn’t leap to who. It establishes what, where, and when: the sequence of events, where people and objects were positioned, points of entry and exit, and the timeline. The most reliable tools are the ones with real scientific foundations:

DNA — the gold standard. A national scientific review found that DNA from a single person, and simple mixtures of two, rest on solid foundational validity. It can place a known person’s biological material at a scene with extraordinary discrimination (though how it got there, and when, still require interpretation).
Latent fingerprints — also found to be foundationally valid when done carefully, with documented error rates.
Forensic toxicology — what substances were in a body, often decisive in poisoning and impairment cases.
Forensic pathology — the autopsy establishes cause and manner of death, the nature of injuries, and often a defensive or offensive sequence.
Trajectory and geometry — the physics of where a shot came from, or how objects moved, can be reconstructed with genuine rigor.
Time-of-death science — body cooling (algor mortis), settling of blood (livor mortis), stiffening (rigor mortis), decomposition, and insect activity (forensic entomology) together bracket when.

Read together, these can rule scenarios in and out — which is often more powerful than any single “smoking gun.” A timeline that makes a claimed sequence physically impossible can break a case open without ever naming a suspect.

Where scene-reading has gone wrong

Here is the part responsible forensics insists on. In 2009 the National Academy of Sciences examined the field and found that several long-trusted “pattern” disciplines were built on subjective interpretation rather than tested science. Its language about one of them, bloodstain pattern analysis, was blunt: the uncertainties are “enormous,” and analysts’ opinions are “more subjective than scientific.” A later large study found bloodstain analysts’ conclusions were frequently erroneous and that examiners often disagreed with one another.

In 2016 a panel of the President’s Council of Advisors on Science and Technology (PCAST) went further, asking which methods had demonstrated foundational validity. It concluded that single-source and simple-mixture DNA and latent fingerprints met the bar — and that bite-mark analysis did not, finding examiners couldn’t even reliably agree whether an injury was a human bite, let alone match it to a person. Hair microscopy and several impression comparisons drew similar criticism. These weren’t academic quibbles: bite-mark and other overstated comparisons have contributed to real wrongful convictions, later overturned by DNA.

The lesson isn’t that forensics is fake. It’s that confidence must be calibrated to the actual science — and that a scene “read” with more certainty than the method supports is exactly how an innocent person ends up convicted.

How we read a scene

We treat a crime scene the way the evidence demands: as a source of what happened, reconstructed from the parts that are scientifically sound, with the uncertainty stated plainly. We separate the validated from the contested. We never let a pattern “reading” harden into an accusation of a specific person — because the scene establishes events, not identities, and because overconfidence here has a body count of its own: the wrongly imprisoned. Understanding how the evidence speaks is the goal; putting words in its mouth is the thing we refuse to do.

Frequently asked questions

What is Locard’s exchange principle? The foundational idea of forensic science, from criminalist Edmond Locard: “every contact leaves a trace.” When two objects touch, each transfers material to the other — which is why physical evidence can document an encounter no one witnessed.

What can a crime scene actually prove? Reliably, it can reconstruct what happened — the sequence of events, positions, entry and exit, and the timeline — and can rule scenarios in or out. The most scientifically validated tools are single-source DNA, simple DNA mixtures, latent fingerprints, toxicology, and forensic pathology.

Which forensic methods have been criticized as unreliable? Major scientific reviews (the 2009 NAS report and the 2016 PCAST report) found that bite-mark analysis, bloodstain pattern analysis, hair microscopy, and some impression comparisons rest heavily on subjective judgment and have contributed to wrongful convictions. DNA and latent fingerprints fared far better.

Can a crime scene identify the killer? A scene establishes events, not identities. DNA or fingerprints can place a known person’s material there, but how and when it arrived still requires interpretation — and treating a scene “reading” as proof of a specific person is precisely where miscarriages of justice happen.

The honest version of forensic science is more useful than the dramatized one — because it tells you not just what the evidence shows, but how sure you’re allowed to be. That humility is what protects the innocent.