Ask a modern forensic scientist whether two samples "match" and you may get a more careful answer than yes or no. Increasingly the field expresses the strength of evidence as a likelihood ratio (LR) — a single number that captures how much the evidence should shift your belief.
The idea
A likelihood ratio compares how probable the evidence is under two competing explanations:
LR = (probability of the evidence if the prosecution's proposition is true) ÷ (probability of the evidence if the defence's proposition is true)
- An LR greater than 1 means the evidence supports the prosecution's proposition; the larger the number, the stronger the support.
- An LR of 1 means the evidence is neutral — equally expected either way.
- An LR less than 1 means it supports the defence.
Laboratories often translate the number into words — "moderate", "strong", "very strong" support — on a defined verbal scale.
Why it's better
The LR keeps the scientist in their proper lane. It states the strength of the evidence, not the probability that the defendant is guilty — which depends on all the other evidence and is for the court to decide. That separation is exactly what avoids the prosecutor's fallacy. It's now standard in DNA mixture interpretation and is spreading across other disciplines, because it is transparent about assumptions and about what the science can and cannot say.
The catch
An LR is only as good as the two propositions and the data behind it. Choosing appropriate propositions, and having reliable frequency data, is where the expertise — and the courtroom scrutiny — lies.