Analysis of the Latent before Comparison (DVI)

[JohnT]

I hope to pose a question that will lead to some healthy debate/discussion. (Of course this may be something that everyone is already comfortable with so there will be no discussion.)

In recent years there has been a strong push for latent print examiners to carry out a full analysis of the latent print before attempting any comparison with the exemplar print (10-print). The push has also seemed to be that they must show that they have done this through notes, methodology etc.

The main reason seems to be to avoid any potential confirmational bias. I personally am not convinced that this is the issue it has been made out to be.

(I'm not questioning analysis to establish, finger, pattern type, target group for comparison etc.)

However the question I pose is this;

In a Disaster Victim Identification (DVI), what print do you analysis first? The Post-Mortem prints (deceased prints) or the Ante-Mortem fingerprints (prints of the victim before death)? Depending on the state of the body the deceased prints may be high quality 10-prints or the equivalent to partial latents. The same can be said for the Ante-Mortem prints. They could be a full set of finger and palms from some sort of police/official record or a latent print from the home/belonging of the victim.

Maybe the approach is to analysis the print with the least quality/quantity of ridge details first before attempting a comparison? But many times in a DVI Operation both prints are poor quality.

My opinion is that the heavy emphasis for analysis (including marking points/making notes) before any comparison is carried out is an unnecessary construct, particularly in relation to confirmational bias, the comparison and evaluation of the prints in question are what decides whether or not the prints are the same, not an isolated analysis of a single print.

[JohnT]

Sorry didn't correct my typos for 'analysis' and 'analyse'.

[Pat]

The main reason seems to be to avoid any potential confirmational bias. I personally am not convinced that this is the issue it has been made out to be.

Hi John

In examinations of terrible, no good, very bad latent prints, I think it is possible for an incautious examiner to talk himself into an identification when he launches straight into the comparison without first doing an analysis of the latent. I believe if you looked at erroneous identifications down through the decades, you would find instances of that in many of them. That said, I also believe the actual number of times that has happened is very small and the ratio of erroneous identifications to the total number of identifications made would be much smaller than one in a million. Those figures, of course, are purely subjective and without any statistical information to back them up. Just my personal belief. And they go not to counter your statement quoted above, but to support it.

The question of whether to analyze the ante-mortem inked print or the post-mortem inked print might not be the correct the question. I believe the best approach would be to analyze the least clear (most distorted, most unclear) print first. Then compare it to the clear print. If both prints are unclear (say, a post-mortem print from a decomposed body to a latent print recovered from items in the victim's home), then perhaps a full analysis of each print would allow a more scientifically pure comparison.

But what we are talking about here is not crime, trial, and conviction. You have a lot of grieving families down in Christchurch who need closure sooner rather than later. We are talking about letting families put their loved ones to rest so they can draw a line beneath the tragedy and move on with their lives. In this situation, I say slap the two prints down side by side and get on with the comparisons. There may be a few really bad prints or unusual circumstances in which a more cautious approach would be warranted, but for the most part, I see no reason to treat those cases the same as criminal investigations in which time is not necessarily of the essence.

[josher89]

John,

I teach a beginning fingerprint class to grad students in Nebraska. After I gave the students a quiz, I was asked by a student why you always look at a latent print prior to the known when doing comparisons (you know, the analysis should always be on the latent first, then the known). I explained this to him and pointed out in Ashbaugh's book (which is required text for my class) on page 108 [first paragraph] that the latent print is always analzyed first. He counted with page 137 [half-way down] where Ashbaugh said you should compare the poorest quality is compared to the best quality. I couldn't argue the quiz question the way I worded it, but in general, latents are usually of less quality than exemplars. So, for my two cents, I would agree that if the poorer quality should be analyzed first because, what would you compare with if your least quality print wasn't comparable in the first place?

[Bill Schade]

I was trained to always work from the latent to the known. This was pre Daubert and the emphasis on the scientific process of ACE-V.
It was explained that by finding detail in the latent prior to looking for it in the known you were eliminating the possibility of "finding" matching points that were not there (bias)

I still believe that is a safe course of action in most instances

HOWEVER

I saw something presented by an eminent fingerprint person in 2002 that made me rethink some of the dogma I grew up with. The issue was that fingerprint comparison was a back and forth process (something I was taught to never admit) and that by only looking at the unknown during the analysis phase, you were limiting yourself to 25% of the information available. The formula below made this premise clear to me and changed my thinking. I now acknowledge that I work "both ways" from unknown to known and known to unknown.

A + C = E

A(50%) + C(50%) = E(100%)

A[25% + 25%] + C(50%) = E(100%)

A(25% + 0%) + C(0%) = E(25%)

"therefore, establishing a predetermined number of points to claim a latent print of value during the analysis phase is non-relevant in a scientific sense."

[JohnT]

Thanks for the comments so far.

Pat - re Christchurch - basically that is what we have done - with superb results. (And the Christchurch FP team are in fine spirits all things considered)

My argument is that the analysing of a latent first (or the print with the poorest quality) has been taken too far. I personally am greatly concern by the idea of making large amounts of notes before commencing a comparison. I don't think it is necessary. What has happened to the days of using a fingerprint glass, doing a simple analysis with my eye and brain and then carrying out a comparison. Any potential error caused by being fooled/swayed by the better quality print shows a lack of understanding during the evaluation stage. In any identification I think the process should be the same. In an effort to get or maintain a label of 'scientific' I think we might be creating processes that are unnecessary.

I have used the example of DVI because of my involvement in a number of DVI operations and identifying deceased persons in general but this relates to all fingerprint identifications.

Thanks for the other comments as well. Bill - I agree. As I typed my intial comments I was actually thinking of the eminent fingerprint person and the formula you quoted. Well done, you read my mind (Bill and I did not prepare this in advance).

[Neville]

Hi Bill
I guess I was never taught this one way dogma I only ever was taught the back and forth method, until about 15 years ago when it was pointed out to me by a trainee, or maybe it was John T, I can't remember, but it came as a bolt of lighting from left field, I just could not understand why you would cut out half of the available information during a comparison. I did try it out for about a couple of weeks and decided it was too confusing, I lost track of where I was, it took 3 times longer, it made no difference to the end result, I suffered the embarrassment of being way down on my idents for the period, I was sure I was the only person in the office doing it and I am sure there where idents I must have missed during that 2 weeks. My little exercise of futility came to an end when a homicide landed on my desk.