With regards to verifying a person's signature, this book states (page 429)

The error rate was 6.5 percent for the experts and 38.3 percent for nonexperts.

What exactly does the error rate represent? If a subject where to claim 2 samples were from the same person when they are not, does this increase the error rate? If a subject where to claim 2 samples were not from the same person when they are, does this increase the error rate?

I found this paper but it still seems ambiguous.

closed as off-topic by Sklivvz Dec 9 '15 at 14:32

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "Skeptics Stack Exchange is for challenging unreferenced notable claims, pseudoscience and biased results. This question might not challenge a claim, or the claim identified might not be notable." – Sklivvz
If this question can be reworded to fit the rules in the help center, please edit the question.


In that paper, they use two measures explained on page 14:

  • Hit Rate - percentage of actual matches that were identified by the subject (i.e. 1 - (type II error rate)), and
  • Wrong association rate (w.a.r.) - percentage of "found" matches that were incorrect (did not actually come from the same person, or, the type I error rate).

It was the w.a.r. that was found to be 38.3% vs 6.5%.

They report both types of errors in the charts, but focus on w.a.r. as "error rate" in the text because that type of error is "much more dangerous because it may link a person to an incriminating document thats/he did not create".

In the book they make this distinction as well, but use the phrases "when false signatures were declared authentic" (equivalent to the w.a.r.) and "when authentic signatures were declared not genuine" (equivalent to 1 - (hit rate)).

Not the answer you're looking for? Browse other questions tagged .