CLPEX Chat Board

Hi,

I am looking for information on the format and composition of comparison tests for fingerprint examiners.
My own force require trainee fingerprint examiners to successfully complete a comparison test as part of their final assessment to become a fingerprint expert.
The test currently being used consists of 10 latents which must be compared with 10 sets of fingerprints. The time allowed for the test is 6 hours. Some of the latents can be identified with the sets of fingerprints, and some cannot. The trainee must reach the correct conclusion on every latent (identified or not identified) to successfully complete the assessment. A 'missed' or eroneous identification will result in the trainee failing the assessment. Though the latents are challenging, all have been carefully selected so that the trainee should be able to reach a decision of identity or exclusion on each one.
Before the assessment begins the trainee is required to complete a brief analysis of each latent, stating it's pattern type (if revealed), likely finger choices and the reasons for those finger choices and any 'red flags' they may observe.
Trainee's are given five latents before lunch (with all the sets of prints) and the other five latents after lunch. The test is split like this to ensure that the trainee's conclusions are their own.

I would be very interested to hear the details of any other tests or proficiency assessments anyone else may have sat or delivered and how they differ from the above.

thanks

Dan

A hundred years ago (well, almost) I prepared latent tests for our civilian fingerprint technicans and our Scenes of Crime officers. Each test consisted of ten latent finger (or palm) prints and twenty ten-prints from our 'known' offenders. The latents ranged in quality, but none were too difficult for a reasonably experienced examiner to identify. There was one 'non-ident' in each packet. These tests were used to establish the level of competence of our current members, and were given to new members as they finished their training.

When I assumed my current duties and began testifying in court, 'testing' became one of the challenges that defense counsel woud raise in an effort to question my status as an "expert" witness. The fact that I was not 'tested' on a regular basis meant that my expertise may have diminished to the point where I was no longer qualified to offer expert testimony. The only counter-arguement I had was the fact that every identification had to be verified by another qualified member of the Unit meant that, in effect, I was "tested" every day. The defense countered with the fact that this verification came from "one of my friends in the office" instead of a disinterested third party. The defense had the same objection to the 'in-house' test that I took. (The fact that I set the tests myself probably didn't help, either)

At that point I began casting about for a disinterested third party and I found Collaborative Testing Services* (http://www.collaborativetesting.com), a company out of Virginia that offers all kinds of forensic testing. The fingerprint test that they sent consisted of five latents and (I think) five sets of known prints (victims and suspects). The latents were of varying quality, but none were too difficult to individualize. (I couldn't tell you how much it cost - I leave that kind of detail to the Administration office).

I passed the test (phew), and now I can state that I have been tested by a disinterested third party that specializes in forensic testing.


*this is not a paid endorsement

In my agency, we have internal and external proficiency testing. Our external proficiency test comes from CTS, usually consisting of 10-11 latents, and 5 standards (finger and palm prints). It is a mixture of ID's and non-ID's. The test is changed annually. I don't take the test every year - I take it every 3rd year, as we rotate who takes the external proficiency test.

Our internal proficiency test is taken annually by every examiner in the agency. It mimics the CTS test in format, but it is created by the examiner who is doing quality reviews. That examiner must also perform case file reviews of 5 cases per examiner every year and random reanalysis of two cases per examiner per year.

I think this is a pretty decent quality review system. It takes into account regular working conditions with the case file reviews and random reanalysis, as well as testing for obvious comparison problems.

People’s postings talk about giving examiners sets of fingerprints to compared. My question is: How difficult are these sets of prints? How is this difficulty determined and quantified? And who decides/chooses (and how) which set of prints to use for the test?

I ask this, as it seems that a test which requires comparing fingerprints highly depends on how difficult they are; they can be trivially easy and they can be terribly difficult. So, the questions above seem to me to be quite critical.

Itiel

Dr. Dror:

It will be interesting to see the different descriptions of difficulty. I've worked through one department's internal tests. For perspective, I've been doing examinations for more than fifteen years, but having played a dual role in small agencies during that time, my experience has been nothing like that of someone working as a full-time examiner. I mention that, because it has a real bearing on how difficult a timed test is found to be.

I would characterize both sets as definitely non-trivial. The test images generally did not include many clues that would make it easy to narrow them to particular rolled print cards. They all had the characteristics of typical latent prints of the sort found in casework. Each test image was of sufficient quality and had sufficient quantity of detail to be doable, but many required considerable experience and aptitude to locate potential matches among the record prints that someone with experience and native ability below a threshold could not work fast enough if they could not efficiently narrow the possibilities. They would run out of time. In other words, anyone with proper training could, given time, correctly identify each test image or determine that it could be excluded from the available record prints. There were no "tricks," no intentionally deceptive images, but there were certainly opportunities to make erroneous identifications. They also included such things as prints captured on, for instance, a rubber lifter that required a mental reversal to compare. There were marked "RUBBER LIFTER," which would clearly prompt any trained person to compare them as reversed impressions.

When time-limited, such tests do not reflect and are not intended to reflect actual casework in which any reasonable time is available and in which the examiner may set a print aside and return to it later with fresh perspective. However, imposing a time limit does to a degree reflect the reality that anyone working in a busy unit. (Are there any units that aren't?) There's going to be a general time pressure during work, so a time limit can tend to detect the inefficient worker who may take chances to meet expectations. But success demands sufficient training and experience that success demonstrates the ability to work efficiently and accurately. It does not, of course, insure that the examiner routinely proceeds daily with proper attitude and rigorous care or will maintain a proper tolerance and a properly conservative view of their own abilities with complex impressions and unusual presentations. It can no more do that than a difficult driving test could insure that the person would not take inappropriate risks and would never drive after drinking or when exhausted.

Dr Dror said:
"My question is: How difficult are these sets of prints? How is this difficulty determined and quantified? And who decides/chooses (and how) which set of prints to use for the test?"

The core claim of fingerprinting is individualisation. A competent fingerprint examiner should therefore be able to routinely distinguish between the real donor of a latent print and the person in the world who has the fingerprint which is the closest match to it. The only way to test this skill would be to include some latent/inked pairs that are extraordinarily close matches but are from different people. I don't know if the people who design the tests take this into account.

“Stress tests” are a normal part of system and personnel testing in many industries. If I was a passenger in a plane flying through mountains at night I would not be happy to hear that the pilot had only been checked flying over Norfolk in sunshine.

Based on my experience, most of our tests (internal and external) range from the trivially easy to the difficult. I usually find a couple of confidence builders, several that require rigorous searching, and one or two that require significantly more time. Of course, this varies with each test. Sometimes the whole test seems easy. Other times it seems excessively difficult.

The problem with creating these tests is that it is not terribly easy to find similar sets of fingerprints within your circle of acquaintances.

As is the case in most things associated with latent print comparison, the discussion about the difficulty of a latent print comparison test tends to be oversimplified. Many factors are of concern with determining the difficulty of a particular comparison and ultimately the difficulty of a test. Difficulty is a function of factors related to the latent prints such as: amount and type of area; type of prints; type of area; background and amount of clarity; amount of quality as in the amount of level one detail and the amount of level two detail and/or level three detail; and types and amounts of distortions. Also the clarity and quality of known prints is to be considered. The number of comparisons time and the time allotted to complete the test are also factors. There are other factors to be considered such as the objective of the test; is the test to be a measure of proficiency or a test of competence? A test that may be easy for one person may be extremely hard for another examiner. A novice may rate a test as difficult while another examiner may think the test is easy. The difficulty of a test is ultimately a function of the experience, training and skill level of the examiner being tested.

Dr Dror,

the marks that are used in the test I described were selected by an expert, then approved by a number of other experts, then put into a test which was taken by two other experts who were close to the level of experience of the people who would eventually be taking the test for real.
The marks were all deemed to be 'challenging' but appropriate for use in the assessment.

I enjoyed your seminar in Manchester at the end of last year by the way - still can't quite believe I didn't see the gorilla.


I'd like to thank everyone else who posted a reply to my enquiry, your answers were much appreciated.

Dan