reporting 'potential' suspects

[Michele]

http://www.newschannel5.com/story/14742 ... t-database

Hendersonville Police Testing FBI Fingerprint Database
Posted: May 27, 2011 4:36 PM PDT

HENDERSONVILLE, Tenn. - The Hendersonville Police Department is trying something new to help make neighborhoods safe. Hendersonville is one of fewer than a dozen departments have been chosen to test an FBI finger print database.

Over the years the art of lifting a finger print has changed very little. Police still dust the surface and lift the print. The end goal is still the same: find the person it belongs to.

"It's like taking two road maps. You take them side by side and you have your known standard and then your latent lift and you compare them and literally it's like taking two transparencies and putting them on top of each other," said Detective David Harrell.

For years, Detective Harrell has sent finger print evidence to the TBI. The TBI compares the prints to others in a data base and sends back the results.

Harrell said the process works, but it's a little slow.

"Unfortunately, they have such a back log that by the time we submit, it's a great deal of time before we get responses back," Harrell said.

Investigators just started participating in an FBI pilot program designed to speed up the process. Investigators send fingerprints directly to the FBI.

The prints are compared to the fingertips of 68 million people in the agency's data base.

"We can get that information back a lot faster than the TBI can provide," Harrell added.

In one case it was just 15 minutes.

While the TBI provides an actual suspect, the FBI sends the investigator a list of potential suspects. It's up to Harrell to analyze the FBI results and see if he can get a match.

"One person on the list may be dead in 1932. One may be in the penitentiary in Washington, DC and the other one may be living out on Elm Street," Harrell went on to say.

But that might be the suspect they're looking for.

It's a little more work, but the potential is there to solve crimes faster and help keep neighborhoods safe.

The FBI pilot program is designed for smaller police departments who don't have their own latent examiners.

The Metro Police Department has examiners who compare prints in the city's data base. Metro also sends prints to the FBI, but only on an as-needed basis.

[Bill Schade]

This is so wrong in so many ways, I can't even begin to express my anger.

If this story is correct I want to know who came up with this plan at the Bureau? It can't be anyone with knowledge of the process of latent print identification.

It has to be a nameless bureaucrat who will never take responsibility for the liability this creates.

GGRRRRR!

[Tazman]

Right, Investigator Harrell! Here's the candidate list of ten suspects we generated from the IAFIS search of the latent print you faxed us. Go forth and arrest these suspects. Waterboard them, if necessary, to get the confession you need. And keep up the good work. We'll have the streets of Hendersonville free of criminals in no time.

[Neville]

Please tell me this is a joke!?!?@?!?#@?!?@##$.

I hope this FBI pilot program designed to speed up the process has been trashed along with the detective. What an idiot (Oxford dictionary) a person to deficient in mind to be capable of rational conduct ( or is that thought)

[Dennis Degler]

[kevin]

Holy cow! The population of this town in Tennessee is nearly 50,000 people. The police department (website link above) has their own mounted patrol, their own diving team, their own SWAT team, their own K-9 unit, motor officers, bike officers, even a cool (and expensive) mobile command center . . . why can't they afford ONE LATENT PRINT EXAMINER?

They are probably saving up for a DNA lab....

[Gerald Clough]

I took some time thinking about this before posting, on account of the apparent dissonance between an FBI pilot program and the vaporings posted in response to the story. Poking around the Web, I did not find anything in general about the "program", so I'll take one story's account of "fewer than a dozen" agencies involved as true. And apparently the Henderson County Sheriff's Office (same county as Hendersonville) began their involvement about a year earlier. The story on the SO program cites four weeks training for operators, which I will assume is limited to AFIS operator training. Neither agency appears to have LPE's from what I can see on their web sites. Based on that, so we have a common basis for discussion, I'll presume that they have installed PC AFIS with operators and are submitting crime scene latents (we would presume they may also use it for checking identity for persons actually in custody) and using an AFIS result as a potential source of suspects.

I want to view this in the same context as other means of producing suspects. There are, of course, many ways. An investigator, especially in a small jurisdiction, may gain a suspect list from the manner of the crime. I have many times put one or more individuals on the suspect list on the basis of the how, what, where, and when of the crime. One might consider an individual as a suspect based on some physical attribute described by a witness. Persons who were known and immediately associated with the description may later be suspected when seen to match. Being found to operate a particular description of vehicle, being seen on foot at an odd hour consistent with the crime, having just been released from jail, or known to have some intimate knowledge of the target of a crime may all create a suspect. Or being on the list of persons who bought a particular model and size of shoe locally, or who purchased a particular model of new tire. None of those thing, of course, can alone produce probable cause to arrest. They merely may prompt some investigation. During the course of that investigation, many individuals may cease to be suspect. They may have been elsewhere, or known to have been local but shown to be not near the crime scene, or to have disposed of the matching vehicle six months earlier, or any number of ways.

There are, of course, always going to be many thousands of individuals around the state or nation that also fit the description or circumstance that led to suspicion. In most of the situations like that above, candidates for suspicion are largely by default locals. Learning that an individual of that description who drives a car like that and who was released from jail in Omaha for the same crime is not going to put that one on the suspect list in Houston, even if the investigator learned of him. We should note that, in what we presume from the news account is the operation in N.C., having no known association with the state pretty much causes an AFIS candidate to be left off the suspect list. It's backward from the usual way, but it operates the same qualification as the more mundane examples. Although they don't discuss it, the nature of the crime may well cause an individual on the AFIS list who has no history of that class of crime to be placed not very high on the list or dropped off quickly. The local pharmacist who happened to catch a DWI ten year earlier is not much of a suspect for the burglary in which two pair of athletic shoes were taken from Walmart. Not completely written off, but not high.

How does AFIS differ from the other, older ways of becoming suspect? How is it like those? AFIS does identify individuals it knows about, generally known criminals, whose physical features (friction ridges) match in a defined way. Like physical description, the match may be semantically correct, but the distinct differences readily apparent. The witness who saw the criminal actor may instantly eliminate a suspect who is a perfect match to the verbal description. Even a fairly naive observer might eliminate many AFIS candidates, and a trained AFIS operator would likely do even better.

Also in common, none of the above, including AFIS, can positively identify anyone. (AFIS is becoming highly reliable for prints taken under controlled conditions, but must be presumed far less so for many latents.) It can only make suggestions. And a bit of inquiry may well indicate none of the suggestions are very good. In investigating many types of crime, my inquiry about who has a particularly described car will be largely limited to the criminal milieu. I'm unlikely to spend much time scouting the no-kids-allowed retirement village for the Walmart shoe burglar's car. There is one question about similarity and difference between an AFIS originating list and other kinds. That is: How likely is it that a local individual will appear on each list? For the conventional sorts of indications, there will certainly be a number of innocent locals who qualify. I do not know how likely it might be that a local will appear on the AFIS list without being the source of the latent, but I suspect that it's generally not very likely. (I suppose that if you limited to the local database and loosened up the tolerance enough, you could get something.)

Would there be investigations where, presuming no available LPE, I would like to know if AFIS selects a reasonable suspect? Certainly. Conventional "legwork" may well prove up an AFIS suspect as the actual actor in time to recover property and prevent loss of evidence. If no such proof can be established, well then, we're just going to have wait for an examination to see if it's him. But, especially in small jurisdictions, you give me a suspect who happens to be the actual crook, and I have a fair chance of working the case to proof before I could get a fingerprint match weeks later. I would also like the results of computer facial matching of store video to mug shots, when that is available. (The technology is pretty good, just rather expensive.) A facial match also wouldn't be probable cause, but if the matched individual is otherwise a reasonable suspect by virtue of locality and other factors, it's a worthy lead.

There is only one unavoidable liability. That is if AFIS produces a very reasonable sounding suspect upon whom you spend considerable time, only to prove he didn't do it. There are, of course, avoidable liabilities. One is the obvious, that an investigator takes the AFIS candidacy as something it's not. But if you imagine that detectives will begin running out and arresting people who appear on the AFIS list, you're being as boorishly insulting as the fingerprint critics who claim examiners are biased toward making ID's and lack the discipline to make reliable judgments of identity. (Sure, it will happen somewhere, just as a doofus examiner will somewhere make a grossly improper identification. Neither is a reason to throw out a resource.)

If the news account is accurate in describing this as a "pilot program," it will eventually be evaluated, which can only really be on the basis of how often the AFIS results lead to an investigation that reaches as least probable cause to arrest the actor in substantially less time than it would take to get an LPE report back, assuming the print is identifiable. Out of the few comments, including jeers, no one said just why they thought it a terrible idea. Is there a rational reason?

I note the proliferation of mobile AFIS. Granted, it's taking live scans under reasonably good conditions, which is pretty reliable, but it is still AFIS suggesting an identity, and it's role is clearly understood to be an indication that more investigation is warranted, either to confirm a given identity or discover a false identify and fugitive status. No more a reflex arrest based solely on AFIS than the described program.

[Neville]

Hi Gerald

You seem to be writing like a detective not as a fingrprint expert.

So if there is a rogue cop who has raped a girl and AFIS produces a list and your name is on it because your prints are there as normal procedures to eliminate police. Do I take it that you will be happy to be suspended without pay while the detectives compile evidence against you to match the offence, until an expert has a chance to confirm the fact that AFIS has not identified you as the offender or even as a suspect. Or perhaps because your name comes up on the list the one untrained fingerprint expert of a neighbouring small town sees that it is you and has always thought you were a bit odd and he confirms the identification due to confirmation bias....

From what I understand the mobile units have inbuilt filters they only come up with a name once a certain score is reached like 10,000 same as live scan and they do not come up with a list of candidates

I still feel that this is a very dangerous practice; I do understand the detective doing what he is doing through ignorance but really! AFIS is not coming up with a suspect list it is producing a candidate list for checking to see if there is a suspect to be checked out by the detective.

Why not just have AFIS produce a list of the offenders and the detective can choose which one he does not like and put him in the electric chair. That would save even more time and money.

[Tazman]

Is the good old U.S. of A. so short of fingerprint examiners that there simply are not enough to check the latents against the candidate lists before the detectives go jacking everybody up?

I mean, come on, even if Gerald is right, wouldn't it save a helluva lot of time just to have an LPE check the candidates? How long does it take to check 15 or 25 candidates? A couple of minutes? If the detective takes the list as a "suspect list," how many man hours of work does it take for him to check out all the suspects?

Wouldn't these agencies be better off hiring a latent print examiner? They would save time and money, and get it right to boot.

I fail to understand this program.

[Kossman]

I think they'd have better odds catching the right guy by just rounding up a few of the "usual suspects" like they did in that movie: "Casablanca". Or maybe just picking a few names out of the phonebook at random.