Randall's Post from the Maryland Thread

[Steve Everist]

Judge Sounder states in her written opinion that “Mr. Meagher incredibly testified that there is no error rate in ACE-V as it is an infallible methodology. He attributed all erroneous identifications to examiner error in applying the methodology. Mr. Meagher was neither credible nor persuasive in this regard. Without impartial testing, however, whether or not the methodology is infallible is unknown.” The methodology does not have an error rate, in this regard, Mr. Meagher is correct. A methodology is merely a series of steps to achieve a particular end. A methodology by its very definition cannot have an error rate (which is not the same as having an error rate of 0%). The methodology does not have an error rate because the methodology, if followed correctly, will always have the same outcome. Does a recipe have an error rate? No, if the ingredients are measured correctly and mixed in the correct order, then cooked at the correct temperature for the correct time, a particular dish will result. If the recipe is not followed correctly by the cook, we have a bad dish and a bad cook.

The ACE-V methodology is used in the examination of friction ridge prints, inked and latent, complete (sic) and partial, and has been accepted as the methodology for this purpose.

In the examination of latent prints, the interpretation of the outcome is the actual question. This is the purview of the examiner. As in the Llera Plaza opinion by Judge Louis Pollack, the training and experience of the examiner are the real question. Certified laboratories have the ability to document the number of comparisons conducted, the number of correct identifications and the number of erroneous identifications, if any, that have been made. Calculations of the error rate of the examiners within the Laboratory could be calculated.

For litigation purposes, the methodologies used by experts to reach conclusions are required by law to be reliable (not infallible). Also experts are required to apply the methodology reliably, but those experts are not required to be infallible. Previously in this written decision, Judge Souder actually stated that examination results need not be infallible to meet the standard for due process. As fingerprint experts we should not be reluctant to admit that fallibility exists in our population and point out that our error rate is asymptotic to zero.

I decided to pull this out because I think it may have been lost in the thread. In the couple of days before this was posted, I had two examiners (independently) say essentially the same thing to me. I'm just wondering how many others have been thinking along these lines?

[Andrew Schriever]

Steve,

I'm glad you started this thread. Randalls post is one of the most clear and easy to understand ways to describe metholodogy and how error rate applies that I have heard.

I actually printed out the post and passed around to my co-workers, and they agreed. Even those who aren't trained in latent prints.

So to answer your question, I have thought on those lines but haven't been able to articulate it as well as Randall did.

Kudos to you Randall.

[Charles Parker]

Randall's post should become standard fare for any LPE training program.

Thanks to Steve for keeping it up front and Thanks to Randall for posting it.

[clpexco]

Here is another distilled thought regarding the reasons behind the attacks on the science. This is from the November 8 edition of Crime Lab Report USA:

The expectation, however, that scientists recognize the potential for error in their work should not embolden those critics who have a disturbingly warped understanding of science. If errors were truly rampant among crime laboratories across the United States, defense attorneys would seek second opinions as often as possible. But they don't. That's because second opinions are more likely to validate the results of the original experts assigned to these cases. It's much safer to simply attack the "science" because there is no risk of collateral damage.