AFIS and the Inconclusive result

[ren6197]

mgirard- Thank you for sharing your experiences with me, I truly appreciate it. Some fear that if you do report it out, this is exactly what could happen, it not being that person. But in the bigger picture, is it really that big of an issue if it was not that person? You never identified them, you clearly stated it was inconclusive. Isn't that up to the detective how to use that information? Or do you think we are being careless by doing so?

Thanks!

-Reyna

[Graham F]

But in the bigger picture, is it really that big of an issue if it was not that person? You never identified them, you clearly stated it was inconclusive. Isn't that up to the detective how to use that information? Or do you think we are being careless by doing so?

I suppose it could be argued that it could be a big issue for the inconclusive person to be denied their liberty for a while, whilst the issue is trying to be resolved. I fully understand that the actions of the investigator against the person you have named are all out of your control. However, if that investigator "strong arms" the individual, plus any combination of other scenarios that could play out and the person complies and provides a clear set of exemplars leading to a change from inconclusive to negative, could the individual have a civil claim against the department? I realize you are some way removed from this, but it was your actions that started the ball rolling. Would you be named in the action? What happens if the individual refuses to comply and provide a set of prints? How far does the investigator go with the case? What if you have colleagues who sit on the fence and call "inconclusive" many times, rather than make a decision? What if the quality of known exemplars you receive continue to be questionable?
If you have clear SOP's to cover your a... that would be great to keep your feet away from the fire. I think the question has already been answered in this thread, in that the personal details of the inconclusive individual are not published, but again this will be dictated by SOP's.

[Tazman]

To take the question to the extreme, if the result of your comparison with the Number 1 candidate is inconclusive and you are going to report that name to the detective, shouldn't you report all of the names of the people in the database to whom the comparisons would yield inconclusive results? Why stop with just the one? Let the detective work on all the people you can't exclude.

[g.]

Reyna,

I have been reading with interest the discussion. Especially the argument, made by several now, "if the person gets arrested...this may result in a lawsuit..." but most of the posters seem to recognize that it's not the fingerprint person's fault, IF the report clearly says "this is not a positive ID, but the person cannot be excluded as donor". [mgirard's story was great though and a nice counterpoint].

That all said, I have not heard of DNA having ANY of these issues. They routinely (DAILY) report out: the person cannot be excluded as a donor (and may share this profile with 100's, 1000's, 10's of 1000's others). I have seen partial profiles searched in state databases where a list of 10-20 possible candidates are given to the investigator (to Tazman's point: yes they should ALL be given to the investigator if they can't be excluded). I have never heard of any agency sued b/c of a database search and the person was on the list as a possible donor.

Nor have I ever heard of an agency sued where the handwriting expert said that it is "highly probable" that the person is the author. Or a footwear impression where the suspect's shoe cannot be excluded, but bears some, but insufficient, individualizing characteristics Or any other forensic discipline which use less than positive IDs/definitive exclusions....

This is the always the argument that is used: the scenario of the cop that goes off on a limb, ignores what we hope would be a clearly worded report, and just arrests the guy. But at least the reality I know for other forenisc disciplines is that this just doesn't happen (or at least to any degree of frequency we need be concerned with). And the other disciplines ROUTINELY do this. Why do we treat fingerprint evidence so differently?

I enjoy this debate since I get it all the time as a primary reason to NOT use statistics for inconclusive latent prints. The fear that is propagated: the jails will be full of people now who are "associated" with the latent print (not identified to it, but as an example: it is estimated that this suspect and 1 in 100,000 others may share these 6 L2D in this arrangement). As I continue to say.... It hasn't stopped DNA one bit from doing the same thing.

Finally, Reyna. Yes our agency recently experienced this. We had the same debate and scenario you described. We reported "As a result of the AFIS search LPX was compared against the known exemplars of XYZ". The result of this comparison was inconclusive due to the limited quantity and quality of ridge detail in the latent print; XYZ could not be excluded as a possible donor". We did provide the name and DOB of the individual, and I think we also added a note at the bottom of the report: "Please Note, that a positive identification has NOT (in bold) been established in this case". I think we added that just to be extra clear. In 12 years I think this has only happened twice here. I think the analyst even called the officer after the report went out and said, "hey let me explain this one to you...".

g.

[RM1023]

Out of curiosity, did the primary examiner write a report saying it was an identification? Then two other verifiers had reports saying no, it's inconclusive? Then did the primary examiner re-write their report to inconclusive? Did the primary then go back to AFIS and look at the rest of the candidates?

[ren6197]

Thank you all for responding to this post, I am really grateful. =)

To answer some questions and comment on others:

Tazman- Great point! All the candidates on the list were excluded, except this one.

Graham F- I completely understand what you are saying. This inconclusive was to a specific finger of that subject, those "almost ID" ones. He could not be excluded. You make a great point in asking if the individual could have a civil claim against the department, and that was the question to be answered. But you never identified that subject, you reported out your true result, it was inconclusive. Nobody that I have asked could be clear if they could sue us, we even asked a District Attorney.

g- Thanks for your feedback! That is exactly what happened here. There is a "disclaimer" on our report, and it will be followed up with a phone call to explain this not so common result from an AFIS search. You bring up a very valid point that other disciplines do the very same thing, and haven't been in trouble for it.

RM1023- The primary examiner was inconclusive and the verifiying examiner agreed with the inconclusive. All the other candidates that resulted from that search were excluded, and all of the fingers except for one of the subject were also excluded.

Thanks again everyone! I really enjoyed this discussion and am so thrilled that we all can learn from one another through this site. =)

Have a great weekend!

[g.]

Thanks again everyone! I really enjoyed this discussion and am so thrilled that we all can learn from one another through this site. =)

Great point Reyna! It probably should be said much much much more often:

Thanks to Kasey for providing this service and managing the site and thanks to Steve Everist for keeping it running smoothly. It is such a valuable resource
and those guys need to be recognized for their service. And thanks to all the other folks to help with the site or generally disseminate info here (and in other places)
like Sandy Siegel, the late Great Charles Parker, Michele Triplett, Haria Haught, Laura Hutchins, Natasha Wheatley, and probably many more that I don't know of. Anyway thanks
to all the info sharers!

g.

[kevin]

And the other disciplines ROUTINELY do this. Why do we treat fingerprint evidence so differently?

I think we are in a different situation with fingerprints - uniqueness is one of the basis for identification. If you report out inconclusive wouldn't it be plausible to believe a layman with little knowledge of the mechanics of latent examination could go in the wrong direction with an inconclusive and treat it like an ident?

As far as lawsuits go I can't see why as detective wouldn't want to go run with it....and you are probably going to be in the right 9 times out of 10. But #10 is the one I would worry about! And as the old saying goes Nimia familiaritas contemptum parit

[Tazman]

It just makes sense that the poorer the quality of a latent print searched through AFIS, the further down the candidate list the correct identification is likely to be, and the more inconclusives are likely to occur. Has anybody done a validation study of their system to determine how many clear points it takes to reliably anticipate that the identification will show up on the candidate list if the correct person is in the system?

[Kasey Wertheim]

It depends on a lot of other factors than how many points it takes. For example, where - how tightly grouped - whether there are missed points in between - and perhaps most importantly, how clear that area of the known print was (how accurately were the matching minutia 'extracted' by the algorithms in the AFIS system). There is no way to tell all this other info ahead of time, so there is no way to know the likelihood of an AFIS match. And it would differ between different systems even if all conditions were ideal.

[Neville]

Hi Reyna
We had a problem here Downunder where the CIB decided to call us as witnesses in cases where (many years ago) we did as you suggest, we had been sending out reports stating it was Joe Bloggs print but not sufficient for court. They wanted us in the witness box saying it was his print but there was not sufficient detail for court, then they had the prosecutor ask but it is his print, what can you say? No? No matter what you say you are in the hot seat with no way out, the jury has heard you say it is his, end of story. The result was we only then reported positive idents. Much better to let 10 guilty go free than one innocent be convicted.

g.
We all have seen idents(?) on AFIS with high scores which are clearly not the offenders, are you suggesting that some time down the track all AFIS hits should be reported when clearly they are false. No, I doubt you are suggesting this but that is where statistics could lead you if used incorrectly. That is if using the statically data and not fully trained experts to make the call.
I think the issue with DNA is that there are not many lawyers who understand the problems with DNA and they are acting like opossums in the headlights. Like the old saying 'blind them with science or baff..........'

Regards
Neville