CLPEX Chat Board

Like all offices our workloads are increasing and our budget is decreasing. My questions are- How many latent prints get searched per expert per day on average? We have various individuals searching anywhere from 35 to 85 impressions per day using NAFIS. How have your offices dealt with increasing workloads and maintaining quality. My personal fear is a drop in quality of work and an increase in erroneous exclusions. Any thoughts and/or solutions are welcome.

Pet peeve: TAX CUTS!!! MUST HAVE TAX CUTS!!! To heck with government services. MUST HAVE TAX CUTS!!!

Okay, sorry, down off my soap box now.

My ID Unit was faced with out of control backlog a few years ago. A sympathetic prosecutor started subpoenaing the Laboratory Manager when he went to trial with a case that had not been examined. I have no recollection of how the prosecutor came up with the idea to subpoena the Lab Mgr -- That's my story and I'm sticking to it! Funny, though, but within a month of the Lab Mgr having to testify to why we couldn't get all our cases done, the Department found the money to hire another latent print examiner.

Comparing numbers inter-agency may not yield an entirely appropriate conclusion. Different policies, procedures, and software could all figure into the equation. More important to focus on your own operation and how it might be optimized. If you are already working at maximum productivity for the current staff within the constraints of policy, procedures, and labor available, then another person may be the only way to do more work.

A very creative solution to a problem We had a few new positions in the pipeline but they didn't eventuate. We're just about fully optimised, I think. Our workflow is fully digital and we are fortunate to have all our latents going through a 'lights out latents' process before we see the images. We often get a suspect to compare before conducting a NAFIS search. We aim for same day searching for property crime and as close as we can to real time searching for major crime but being in a growth industry it gets harder every year.

I was keen to see if the broader 'fingerprint community' had one or two ideas we hadn't already come up with. If a few people have ideas we might all be able to optimise our processes a little.

The witness box can be a powerfully motivating factor

Hey Aaron,

Does your agency by chance have any criteria on exclusions? Incorporating exclusion criteria may help speed up time spent comparing while also addressing your fear of an increase in erroneous exclusions. Our agency has criteria that in order for an examiner to justify an exclusion decision they must be able to demonstrate and justify the area/orientation of the latent, have a clear target group (free of distortion) with some sort of focal point that narrows the search to a specific location, and performed a thorough search (our recommendation is an average of 30 minutes). In doing this your agency would need an additional decision to address latents that can’t be excluded if you don’t already have one. We word one of our inconclusive decisions as an inability to locate due to ambiguous features. This has significantly cut down on examiners spending weeks on that one elusive print that they just can’t find and is a more honest reflection of the prints we get. Another option would be to have some sort of measure of the complexity of a latent/comparison. If you have that, you could develop quality assurance measures that are based on what is needed. For example, full verification on complex comparisons (more prone to error) vs. a technical review on basic comparisons (less prone to error but could have clerical errors). You could also look into what is being run through your AFIS database. What kind of latents are you keeping of value? We don’t run every latent through AFIS because we keep a lot of borderline latents for comparison. Do you run/compare clear child/officer impressions?
Would you be able to share your agencies SOP? How does your agency handle reporting? Do you provide narrative reports or are they automatically generated?

Also, being levy funded is a great way to get around some budget restrictions. We are very fortunate in to be levy funded. It’s a minimal cost to tax payers but allows us the luxury of hiring the staff needed for our workload (we currently have 18 examiners and I do a couple cases a month as a working supervisor to stay knowledgeable of our unit workflow). We had a backlog for a couple years due to being short-staffed but now we have no backlog and our average turn-around-time has gone from a little over 3 months this time last year to about 2 weeks currently. I don’t know if that is something your agency would be able to consider but it certainly has been nice.

Amanda

Thanks Amanda,
I appreciate your comments and advice. I've sent you a message.

I should clarify my initial questions.

I believe we are already quite efficient, particularly through the use of a fully digital workflow, but I'm sure there are people and offices out there who have achieved small gains in efficiency through clever thinking and application of technology or procedures. Examples I've heard of include, only searching one impression in a multiple finger grouping or searching a set number on impressions for less serious crime, say 10 for example. We currently search all impressions set to us for all crime from graffiti to murder. All serious crime is independently searched twice.

Has anyone's office tried anything, large or small, to obtain little, or large, gains in efficiency (or reduced numbers of impressions to search/compare) to enable experts to spend longer looking at comparisons? If various initiatives have been tried, how did they turn out?

I know this is a bit of a shot in the dark but I'm always interested in how the rest of the world does things. I'm in Australia.

Aaron

Our agency utilizes a "one and done" policy that helps speed things up. All AFIS quality latents are searched, but if a subject is "hit" multiple times only 1 latent would go through the full Identification/verification process. The remaining "hits" to the subject would have the AFIS paperwork printed (markup & side-by-side screenshot), but would not be identified/verified. If a subject just "hit" once that would go through a normal Identification/verification. Additionally, if a subject is generated during an AFIS search, any non-AFIS quality latents in the case are not automatically compared to the subject. Investigators may request the latents be fully worked, but they never do. We do exclude homicides from this policy though and those are fully worked (all "hits" IDed and non-AFIS quality latents compared to generated subjects).

Another policy we are moving towards is changing our suitability options so we are not wasting time marking latents that we may never compare. For example, if there are no suspects and the case is going to be worked as an AFIS case, only AFIS quality latents would be marked. It would be noted and reported if there are comparison quality latents that remain in case, but they would not be marked until requested.
PG

Hi again Aaron,

I'm just curious, how many candidates do you get as a default for NAFIS? Our agency utilizes 3 systems (local, regional, national). For our local database our default is to return 15 candidates (this may get lowered once we upgrade our system). For our regional and national systems we default to 6 because the vast majority of our "hits" occur within the top 3 candidates. I worked previously at an agency that would return 20 candidates as the default for the state database so I know each agency will set a threshold based on their resources.

Some exceptions we have on what is run through our databases are: only 1 print needs to be searched for same source impressions, obvious child and officer prints aren't required to be run, for potential simultaneous impressions all of them get searched locally but only 1 is required to be searched through our external databases (a lot of our examiners still run them all though).

I know some examiners will utilize closed-searches in AFIS for some comparisons to free them up to perform other comparisons. My understanding for the reasoning is that this doubles the amount of searching done concurrently.

Peter,
your agency sounds very similar to mine. We do all those things with the exception of taking a screen shot of the NAFIS hit screen. I might look into that. We record all the ID's/hits but don't report it to the investigator. Our next generation system that is in the pipeline....... one day.... will apparently work with our Forensic Register to save the hits when we mark them up and declare them in NAFIS.

If we get a NAFIS hit we will some times do a closed search on NAFIS to see if we can identify further latents but that is up to the expert doing the search. Usually this would be done if the first identification wasn't good evidence (outside of a window versus the inside of the window etc) Major crime is always fully searched by two experts.

We currently do not mark up latents that are not suitable for NAFIS and this does save us a lot of time. Only prints being searched are coded. If a case is going to court then a full comparison and search of every identifiable latent is conducted prior to submitting the court file. We don't currently indicate all the identifiable prints that are not NAFIS quality and I might bring that up at work as a case file note taking issue.

Thank you for your reply. Its handy to know what others are doing and I'm getting a few extra things to think about along the way.

Aaron

Hi Amanda,
We get a list of up to 100. Usually between 75 and 100 candidates are returned. We don't check them all and most experts check between 5-10 candidates depending on the impression and type of crime. I'm not sure how other AFIS systems are set up but we can send 5 fingers and 3 palms for searching at a time from one case. There can be as many prints in the case as you like but only the 8 can be sent at a time. If NAFIS finds the same candidate in 2 or more of the up to 8 searches it indicates what we calla a multi. Generally a quick scroll down the list reveals the indication or not. I've had candidates identified at around 13 and 40 before. I would not have found them without the multi. It's very handy. They aren't always hits but it helps.

We can choose to search nationally or by state but either way we get the same number of candidates. We run our property crime through our state and the nearest adjoining state and all major crime is run nationally.

I agree that most hits are in the top 3 candidates but I do like having the longer list to search for multis. I just read Ed German's presentation from last years IAI conference and it seems the FBI might also be comparing the top 3 candidates. For 'volume'/property/non-major crime it does seem like a realistic cost benefit scenario. Our office could certainly gain some time by adopting that strategy but it would be looked at cautiously.

We also conduct closed searching when its appropriate and it certainly is quicker than an open search.

Thanks again for your questions and feedback.

Aaron

Point of order.

Ed German's presentation was Presented 30 October 2017 at the Cogent User Group International Educational Conference. Obtained at Onin.com.