CLPEX Chat Board

The thread on Acosta-Roques v. Holder prompted me to post this, since there were some comments on documentation of tenprint-to-tenprint comparisons.

In our office, when we get a comparison request (from, say, a prosecutor or a detective), this is our basic workflow:
1. Perform the comparison
2. Initial and date the particular finger used on idents
3. Type up the report
4. Give the exemplars, with the report, to a second examiner to verify
5. The verifier initials and dates the finger(s) he/she used, signs the report, and gives everything back to the initial examiner, who sends out the report

The possible bias inherent in giving marked-up exemplars and a report showing one's conclusions to the verifier should be pretty obvious, including to defense attorneys. So I wonder if other tenprint shops out there generally just give the prints to the second examiner to do a "blind" verification and then do any initialing/signing of reports after.

I also wonder if other agencies initial the finger used or just initial/date the card or other document showing an ident was made. My sense is that most do the latter. We had a case recently where the defense tried to imply bias in the fact that both tenprint examiners had looked at the same finger. I'm not sure there's any real value in marking the particular finger used, anyway.

Comments?

I was under the impression (no pun intended) that ten print comparisons didn't routinely require a verification. I think the NAS report even stated that ten print to ten print comparisons are reliable enough for computers to perform 'lights out' searches without any major snafu's. That being said, we are required to initial the finger block (of the finger used for the ident) and then we initial and date the bottom of the card. Let's be honest, here, though...there are 19 other fingers to look at as well (rolls and slaps). So, one examiner could look at the right thumb rolls and the verifier could look at the right index finger slaps.

Our workflow is similar to yours in that we perform the comp., intial the finger, type the report and give the exemplars with a case file to the verifier who is also the tech reviewer. We don't mark up the prints (generally) on a ten print to ten print comparison (they're usually pretty obvious) on the card but we may enlarge the prints and include a charted copy in the case file. The verification is usually done prior to the tech review.

It all boils down to whether or not blind verification avoids bias (or potential bias) or if it does nothing to reduce it. I could say I'm being biased (even if you didn't have the finger marked) because you are giving me on person's card, implying that you've identified it to one of those fingers. I've heard of other shops that produce five fingerprint cards with similar henry class. for these types of idents and verifications to avoid "bias"...WAY TOO MUCH OVERKILL AND ABSOLUTELY NOT NECESSARY!!!!

Alex Mankevich wrote a great article in the Chesapeake Examiner that I'm going to try and post here. I hope I'm not violaing any copyright laws by posting; if I am, can a mod please remove the link or advise me? Anyway, he distinguishes between an independently-performed experimentation and independently-replicated conclusion. It's a great read and I'm not going to regurge it here but I would recommend that people read it. I really like the way he explains bias and his analogy to astronomy. I even use that in my fingerprint class I teach.

No matter what, the debate on bias will go on and on but the bottom line is, I don't think that bias would prevent reliable conclusions in print examinations, especially ten print to ten print examinations.

There is a word that defines the condition under which conclusions are independently repeated without access to or knowledge of the original data, observations, and interpretations. That word is 'happenstance.'

Brilliant.

josher89 wrote:I was under the impression (no pun intended) that ten print comparisons didn't routinely require a verification. I think the NAS report even stated that ten print to ten print comparisons are reliable enough for computers to perform 'lights out' searches without any major snafu's."

I'm not referring to AFIS searches, just the miscellaneous other comparisons we do upon request from outside agencies/investigators.

Thanks for the input!