AFIS and the Inconclusive result

[ren6197]

Hi All!

Although this is my first time posting, I have been and avid reader of this board for quite some time. Thanks for all that give their input, it helps out tremendously. =)

Can I please have your opinion on the following scenario:

You enter a latent print into AFIS and get a candidate to compare. In the comparison stage the primary is able to identify it, but the verifier cannot and will have to go inconclusive to that particular finger. The primary drops down to inconclusive. What happens next? Would you put a report out with that candidate's name/info stating that it is an inconclusive result from an AFIS hit? Or do you say "no hit" and move on to the next? If you don't think you should send out a report stating your conclusion as an inconclusive (with the candidate's information), what is your reasoning? Wouldn't it be a great tool for a detective to use to look more into a particular subject? Is it even ethical to do this? Has any other agency sent out an inconclusive report resulting from an AFIS hit?

There are so many other topics within this one topic but those discussions will have to come later. For example, why are we putting in borderline prints into AFIS, sending the comparison out to a 3rd examiner, etc.

Thank you in advance for your insight!

-Reyna

[ER]

I don't think my agency has done this before, and it probably wouldn't come up since we don't search borderline prints through AFIS, but...

If you think that this print belongs to this guy but there's just not quite enough to ID, then it might be best to report the comparison, since you actually did do a comparison. In this case you may want to consider wording similar to what Vegas Metro uses on close inconclusives (JFI 61-3). This basically says that the result is inconclusive, but there is limited detail that is consistent.

I think that the key question here is do you report an inconclusive result from an AFIS search or do you not report a comparison that was actually conducted? Reporting what you did seems like the more transparent choice.

[ren6197]

Thank you for your reply ER.

I guess this question would naturally come next. Why don't you search borderline prints through AFIS? What I mean by borderline is a latent that maybe you would identify, but that others in your unit would probably not. Is there any harm in entering those prints into AFIS, and reporting out an inconclusive with the subject's name/info? Would it be a good lead for a detective? Are we crossing boundaries in our LP community? I understand inconclusives resulting from a comparison case with a name that was given to you, but you actually giving the name to a detective. Is that different? Should it be?

This is a hot topic in our area right now. =)

[ER]

We don't search borderline prints because
a) they usually won't hit, even if the database contains the suspect;
b) an inconclusive result usually isn't useful to the case; and
c) there is an increased risk of error by comparing low q/q prints to millions of fingers

Since we don't search these, then we avoid the debate of what to do if you find the guy but there's not enough to call it.

[Graham F]

What I mean by borderline is a latent that maybe you would identify, but that others in your unit would probably not.

This raises other questions, i.e., do the "others' need more experience, more training, are they sitting on the fence with their decisions, are you over stating your conclusion? Perm any of these with other questions this may raise.

Is there any harm in entering those prints into AFIS,

There is no harm in entering these if you feel you can identify them, as you never know until identified, if your colleague will support your conclusion. Of course garbage in can = garbage out. A sub standard latent can continually respond to all reverse searches and may therefore become such a pain that it will eventually be removed from the system.
You should not curtail your skill level and dumb down to that of your fellow examiners, by not running a latent through the system. You have the option of searching a poor quality latent and if this is negative, the latent can be removed. At least you attempted to locate the donor. Leaving the latent in the case file assists no one, if in your opinion the latent is AFIS searchable.

[PCC]

You dont say why it is inconclusive. When we have encountered this it is because the known print card is lacking. So we generate a report stating that the result is inconclusive and will be re-compared when a better quality known print card becomes available. If the problem is the latent, then we mention that the latent is not of comparison value.

[Tazman]

If you don't think you should send out a report stating your conclusion as an inconclusive (with the candidate's information), what is your reasoning? Wouldn't it be a great tool for a detective to use to look more into a particular subject?

Hi Reyna

There are two kinds of detectives in this world. The first kind would begin gathering background information on the person you named and do his homework before acting on your report. The second would dash out and haul the suspect in for questioning, most likely under arrest on a warrant for which the detective had exagerrated the information he put into the affidavit.

In the second scenario, your department would then have to deal with a false arrest lawsuit.

Better to just report "no identifications were made" and move on.

[Boyd Baumgartner]

I don't think my agency has done this before, and it probably wouldn't come up since we don't search borderline prints through AFIS, but...

Based upon your followup post you say

We don't search borderline prints because
a) they usually won't hit, even if the database contains the suspect;
b) an inconclusive result usually isn't useful to the case; and
c) there is an increased risk of error by comparing low q/q prints to millions of fingers

So the inference here and correct me if I'm wrong, is that value (AFIS value in this regard) is based upon a minimum predefined 'clear' points (c) and that you can say with some probabilistic certainty that things under this threshold won't hit (a) and that you know how the evaluation will be used investigatively and forensically in court (b).

Do you have formal definitions for 'borderline' and 'inconclusive' and if so, how are they worded in your SOP?

[ren6197]

Thank you all for your responses. =)

Graham F- If one examiner enters a low quality/quantity latent print into the database, and hits on a subject, they may feel like they can identify it. There are obvious problems with the latent print if others in the unit cannot, and will have to go inconclusive due to the insufficient quantity/quality of the latent print (assuming those other examiners have the training and experience to back it up). At that point, when your verifier does not agree with you (or even the 3rd examiner for that matter), do you send out an inconclusive report with the subject's name and info on it? Or do you just say there was "no hit" because you could not come to an identification conclusion?

PCC- It would be inconclusive due to the insufficient quantity/quality of the latent print. You can't pull the trigger for an ID and you cannot exclude the subject. The known prints are not lacking, it would be the latent print.

Tazman- Exactly! My personal fear would be your second stated scenario and a false arrest lawsuit. But to push it a little more, would we not be covered in the report because we never identified that person? We clearly state it is an inconclusive and not an ID. Maybe educate the detective on what exactly it means so they know this is strictly for investigation purposes only? Is that me living in la-la land? =)

Thanks again!!

[kevin]

There are obvious problems with the latent print if others in the unit cannot, and will have to go inconclusive due to the insufficient quantity/quality of the latent print (assuming those other examiners have the training and experience to back it up). At that point, when your verifier does not agree with you (or even the 3rd examiner for that matter), do you send out an inconclusive report with the subject's name and info on it? Or do you just say there was "no hit" because you could not come to an identification conclusion?

I think what most agencies do is have a written procedure for examiners that are in disagreement - usually this would involve a third agency or outside person to look at the print.

[ren6197]

Hi Kevin! Thanks for the reply.

The thing is that we do have two examiner's agree with the inconclusive conclusion, the question is now what do we do with it. This is an inconclusive to the latent conclusion from an AFIS search. What do you do with that? Report out a name that you provided from AFIS as an inconclusive? Report no hit? The whole inconclusive from an AFIS hit is new to us, I was wondering if any other agency has seen and dealt with this yet.

Thanks!

Reyna

[kevin]

Hey Reyna,

I get where you're coming from now - I saw where the examiner(s) changed their conclusion and thought that was where you were going with it....

I'd treat AFIS candidates as candidates and nothing more - you search your prints and attempt to find a match but you can't treat each of the AFIS responses as evidence or comparison submissions or anything. So I'd report it as 'no hit' and be done with it.

[mgirard]

The unit in which I work has experienced a few occasions with this sort of AFIS situation. In the most recent case the latent was of medium quality (pattern type obvious, core to delta count obvious, 12-14 minutiae searched, clarity was not great in the core or around the delta) and one of the AFIS candidates showed considerable agreement but some minutia were slightly off but could possibly be explained by distortion and poor clarity in those areas. The analyst entering the latent saw enough agreement to consult another analyst. Both of them felt that because there appeared to be so much agreement and not necessarily enough disagreement that couldn't have been the result of distortion they would have to call it Inconclusive. It ended up that that information was included into the case file but not reported out because at this time our SOPs are to only report Individualizations as a result of AFIS. We don't report out all of the exclusions to candidates we look at as a result of AFIS. I think we are already aware that searching a database with millions of fingerprints you are bound to find some that are somewhat similar, ie. Brandon Mayfield, and may need to be a little more conservative based on the quality of the latent. During Technical Review it turned out that one of the victims in the case had a fingerprint standard on file that the original analyst hadn't realized. Once that comparison was conducted it was determined that the latent could be individualized to the victim and the slight inconsistencies that were noted with the AFIS candidate clearly were in agreement with the victim. So I would be on the side of not reporting out an Inconclusive as the result of AFIS.

[ren6197]

Kevin- Good stuff. =) I know traditionally in our community we have done what you are saying and reported out a no hit, but why wouldn't you send out an inconclusive report with the candidate's name and info? What would be the harm in it? This is assuming your unit reports out inconclusives to the latent print to begin with. We came upon this situation and some feel it is wrong to send out an INC report with an AFIS candidate's name, but there is no concrete reasoning behind it. Is it just because it is something new?

mgirard- This is great info! I wondered if any other unit has come across this same situation and what they have done with it. So the latent print you mentioned was identified to the victim, was the victim your AFIS candidate you went inconclusive on? Everything you brought up are great and valid reasons but the question is still asked, why not? You are not identifying that person, you are clearly stating it is an inconclusive. Why not add the information in a report? Would it not be up to the detective to follow up on that name build probable cause from there? The information would be no means to arrest somebody.

Thanks for your input!

Reyna

[mgirard]

No the AFIS candidate was not the same person as the Victim, just happened to share a lot of similarities with the latent. AFIS was doing what it is meant to do. Again it is our policy to report out only the individualizations made as a result of AFIS. If you don't have a policy then it is up to each analyst to decide what they want to do but that's when disagreement and hard feelings can occur between analysts as well. In all of the instances I have observed where there were similarities noted due to AFIS that could have potentially been misinterpreted by an analyst as more significant agreement then what was actually occurring it has turned out that the print actually belonged to someone else. So based on my experience with this type of situation it would be more harmful than helpful to the case and to the AFIS candidate.