Detail Archives    Discuss This Issue    Subscribe to The Detail Fingerprint News Archive       Search Past Details

G o o d   M o r n i n g !
via THE WEEKLY DETAIL
 
Monday, June 25, 2007

 
The purpose of the Detail is to help keep you informed of the current state of affairs in the latent print community, to provide an avenue to circulate original fingerprint-related articles, and to announce important events as they happen in our field.
_________________________________________
__________________________________________
Breaking NEWz you can UzE...
compiled by Jon Stimac

From Bloody Fingerprints to Nasty Drains  NAPLES DAILY NEWS, FL- Jun 22, 2007 ...crime scene cleaners do their job...

Suspects to be Fingerprinted TRINIDAD NEWS, TRINIDAD & TOBAGO - Jun 21, 2007 ...Police Service Act which will give police the authority to force persons accused of crimes to give their fingerprints...

Fingerprints Don't Match, So Accused Man is Freed   SHEBOYGAN NEWS, WI - Jun 19, 2007 ...a man accused of burglary has been freed after spending more than three months in jail after new fingerprint evidence trumped old footprint evidence...

Solving Crime Focus Of Summer Camp CLICK2HOUSTON.COM - Jun 18, 2007 ...university is helping students get a real-life look at solving crimes during a summer camp...

__________________________________________
Recent CLPEX Posting Activity
Last Week's Board topics containing new posts
Moderated by Steve Everist

The Lockerbie Connection.
Iain McKie 378 Sun Jun 24, 2007 3:59 pm

Latent Print reporting and Inconclusive Determinations
Macgyver130 2439 Sun Jun 24, 2007 2:34 pm

Latent Print Supervisor Position - King County Sheriff
Steve Everist 627 Sun Jun 24, 2007 1:11 pm

Interesting Tidbit 4
Charles Parker 480 Sun Jun 24, 2007 12:28 pm

Contemporaneous Documentation
Michele Triplett 129 Thu Jun 21, 2007 2:43 pm

General Electric v. Joiner
Charles Parker 178 Wed Jun 20, 2007 2:58 pm

Latent Print Examiner Positions - CONUS/OCONUS
wkpetroka 719 Wed Jun 20, 2007 5:47 am

Texas Daubert
Charles Parker 130 Wed Jun 20, 2007 3:31 am

Statistics and Misidentifications - The weeks Detail
Michele Triplett 8170 Sun Jun 17, 2007 7:51 pm

(http://clpex.com/phpBB/viewforum.php?f=2)
 

 
UPDATES ON CLPEX.com


Next week we will be voting on our 2007 CLPEX.com T-shirt slogan!  It's time to get your ideas in for voting.  Over the course of the next week, come up with a good t-shirt slogan and e-mail me at kaseywertheim@aol.com, or post your slogan to the CLPEX.com board (located on the left of the home page).  I think we will be able to have them available in time for the IAI conference in San Diego. 

On that note, unfortunately I found out last week that for the first time since I became a latent print examiner in 1997, I won't be able to attend the IAI conference.  A significant forensic DoD conference exists the same week where I will be briefing latent print examination related activities.  I hope to catch everyone at the 2008 conference in Kentucky.  If you have the chance to meet the two guys we are able to send, they are Tom Kern and Brian O'Hara - they can pass on recent info on our work, our employer, and opportunities that are available at different locations of interest, whether they be sandy or not.  Please reply if you would like me to get you in touch with them by e-mail.


_________________________________________
 

Last week

We continued a series on latent print reporting, the process, conclusions and error.  We discussed milestones along the path of crime scene to the Analysis phase of ACE-V that can negatively affect what would have been a conclusive outcome.

This week

We continue the series with a discussion of the Comparison and Evaluation phases of ACE-V.

_________________________________________
Hypotheses of Latent Print Comparison and Evaluation
by Kasey Wertheim

Remember that if we are in the Comparison phase, we have already concluded during Analysis that the impression is suitable for comparison, and we have already chosen our target group. So what is the first thing that occurs as you begin searching or comparing the target group in your mind against the known prints?

Logically, the first thing you actually consider is the location of the target group within the known prints you are searching. If you chose a group of detail around the right delta from what appears to be a left thumb print, you first begin looking for right delta’s on left thumb prints. Only if you find one do you look for your target group. Although this decision happens within a split second of comparing the target group to an impression, what you have actually done is hypothesized that the known print contains the area that corresponds to the latent print area you need for the identification. If the process of comparison leads you to the conclusion that the known print does not contain the area necessary for comparison, you will most likely conduct another analysis of the latent print, choose a second target group, and conduct another comparison. This process continues until you can prove your hypothesis that the known print contains a suitable or corresponding area for comparison with the latent print. If it does not, then your evaluation would be that an actual comparison to the same area of that finger was impossible, and you would most likely request “complete” known prints so that a comparison could actually occur.

At the same time you are searching for suitable area, you are also analyzing the quality of that area of detail within the known impression. How often have you had a pristine latent print with pore and edge detail only to find that the known prints were so poorly taken that an identification is not even possible? It is only during the analysis of the known prints, usually conducted during comparison of the latent with the known print in the same area, that we realize the known print does not contain suitable quality for comparison with the latent print. Generally, this hypothesis is also proven within a split second during comparison because often we will find our target group even in poorly recorded known prints, and we proceed to whether there is sufficient clarity of agreement for identification. Which brings us to the real issue at hand – what is our next hypothesis after establishing that the known print contains suitable area and quality?

In any scientific experiment there exists a true and false statement, and evidence is gathered to disprove one of the two statements. In his article, "The Scientific Comparison and Identification of Fingerprint Evidence," Pat Wertheim describes the two statements as the hypothesis and counter hypothesis. "If the hypothesis is identification, then the counter-hypothesis would be exclusion." Pat goes on to explain that "the “null hypothesis” might include the possibility that the examiner could not reach a conclusion." In statistics, the term "null hypothesis"(H0) is used in combination with the term "hypothesis" (H1) to represent the two contrasting statements. In an experiment, the hypothesis is true if the probability is very low that the null hypothesis is true. For this reason you want the hypothesis to be what the experiment is trying to prove.

So the question becomes what is the hypothesis that a latent print examiner is trying to prove during the comparison of a target group in the same area of two “sufficient” prints? If we start from the position trying to prove that the target group matches a group of details in the known print, I propose we are setting ourselves up for a pretty hard fall. If we are truly objective in our approach, hypothesizing that the prints match before we even begin the comparison is generally not appropriate. On the other hand, if we try to prove exclusion, that isn’t being entirely objective either. We know that our standard methodology is mostly objective, and that it insures that we begin from an objective starting point. We remain objective throughout the process by conducting comparison from the unclear to the clear. If we incorrectly conducted comparisons from the clear to unclear, we would be asking for the every-day strengths of the human brain for object recognition to be used non-objectively. This could lead to seeing things in the unclear impression that don’t actually exist.

I propose that we start with the assumption that the comparison will be inconclusive, and objectively try to disprove the hypothesis by gathering evidence in the impression that refutes it. If you think about it, don't you force yourself to start the comparison from a neutral stance?  This also falls in line with the psychological aspects of comparative analysis as articulated by David Ashbaugh in Ridgeology. Through a neutral starting position combined with thorough analysis, and comparison from the unclear to the clear impression, we avoid the problem of mind "set" and insure that we remain objective throughout the examination process.

So if we start with the assumption that the comparison will be inconclusive, then that would be our hypothesis (H1) we are trying to "prove".  If we discover during comparison that the chances of an inconclusive determination are so low that they can be disregarded, then we overwhelmingly support our null hypothesis (H0); that we can reach a conclusive determination. The next logical question is, what is our null hypothesis?

The author of "Stating the Hypothesis" explains that the hypothesis and null hypothesis are mutually exclusive (there are no overlapping logical conditions) and exhaustive (all possible logical conditions are covered)." This means there would be no opinion outside of these two hypotheses. But if our hypothesis was as general as “the comparison will be inconclusive”, then the counter hypothesis would have to be just as general to be exhaustive and mutually exclusive: “the comparison will be conclusive.”

I propose instead that we actually use two sets of hypotheses and null hypotheses that lead to four conclusions within the Comparison of two impressions. In reality, we start the comparison phase from a completely neutral perspective - no similarity or dissimilarity has been found - and depending on whether we perceive similarity or dissimilarity, we transition into one or the other hypothesis set. If we first encounter or begin to see similarity, we begin to support our null hypothesis that similarity is present between the latent and known prints.  Therefore we begin to disprove our hypothesis that no similarity is present.  Conversely, if we first encounter or begin to see dissimilarity, we begin to support our null hypothesis that dissimilarity is present between the latent and known prints and therefore begin to disprove our hypothesis that no dissimilarity is present.



Arrows are drawn both ways to represent that we can go back and forth between viewing what we consider similarity and viewing what we consider dissimilarity.  We can also cycle back and forth between analysis and comparison. 

We continue to attempt to prove the hypothesis by looking for a lack of similarity or dissimilarity, but in fact by finding it we continue to support our null hypothesis.  By operating in two sets of hypotheses, we are unbiased and objective as we conduct our experiment.  And from experience teaching courses and speaking in detail with examiners about our process, this in fact does accurately reflect what is going on in our mind during a comparison.

In fact it is sometimes even appropriate to start at the far end of the spectrum. Take for example an examiner who is given a pair of impressions to conduct a verification and they are told that the prints “definitely match” and that a “quick verification” is necessary, it can predispose the examiner toward a conclusion. This is referred to by some as “confirmation bias” that can destroy objectivity in the verification process. However, if the verifier takes the opposite approach and in fact starts the comparison with the opposite hypothesis - that dissimilarity will be found - this bias may be avoided. By spending the additional effort to in fact disprove the first hypothesis and switch hypothesis sets, and then continue in the comparison process to prove the opposite hypothesis - that similarity will be found - the examiner has basically covered additional comparison ground that in fact lends more weight to the conclusion and can have added weight in dismissing preconceived notions. However, in a problem identification scenario, there may not be an abundance of detail and this technique may backfire, resulting in an inconclusive determination rather than the examiner being able to reach a conclusion if bias had not existed. If the comparison of a “tough” ID is started with the hypothesis that the prints contain dissimilarity but enough evidence is not found to cause a switch in hypothesis sets, it would be very difficult for that examiner to begin the comparison in another session from a position of “inconclusive” and actually be comfortable with finding sufficient similarity to establish individualization. Likewise, if the comparison of a “close” exclusion is started with the hypothesis that the prints contain similarity but enough dissimilarity is not found to cause a switch in hypothesis sets, it would be very difficult for that examiner to begin the comparison and find sufficient dissimilarity to exclude. It is therefore best to reduce or eliminate bias from the examination process when possible.

In summary of the Comparison phase of the latent process, we first prove to ourselves that the correct area of the known print is present by finding the area.  We continue very quickly, sometimes almost subconsciously, by proving to ourselves that the area is clear enough to see our target group.  We then shift into whether we see similarity or dissimilarity of our target group until we begin to consider sufficiency.  If we distill down these hypothesis sets and add them to our running series total of requirements for a correct conclusion, we find that these 4 hypothesis sets define the Comparison phase of the methodology:

COMPARISON

19) Hypothesis 3: Known print contains suitable area for comparison with the latent print
20) Hypothesis 4: Known print contains suitable quality for comparison with the latent print
21A) Hypothesis 5a: No similarity is present between the latent and known print
        5aHo: Similarity is present between the latent print and known print
21B) Hypothesis 5b: No dissimilarity is present between the latent and known print
        5bHo: Dissimilarity is present between the latent and known print

When we have gathered bits of similarity or dissimilarity to the point that we are actually starting to wonder whether sufficiency for a conclusion exists, we switch to the evaluation phase of ACE-V.

During evaluation, we weigh the quality and quantity of similarity or dissimilarity we have found to determine if sufficiency for a conclusion has been reached.  This is an individual opinion of the examiner conducting the comparison, and it is based on a host of factors including training, experience, talent, motivation, daily variables, and others.  See Pat Wertheim's Ability Equation for more detail.  Since to each examiner this is their personal decision, it is generally regarded as unprofessional and inappropriate for an examiner to be pressured into a conclusive determination when they are not in fact conclusive.  The individual nature of Evaluation will be critical to consider as we explore errors.  For now, upon consideration of whether sufficient similarity or dissimilarity exists, we again have to start from a position to "prove" an inconclusive determination, and only if we falsify our hypothesis can we declare our own conclusion. 

In the evaluation phase of ACE-V, we again start from the neutral position neither of agreement nor disagreement.  If it can be disproved that sufficient similarity does not exist to individualize, then there is sufficient similarity (agreement) and therefore individualization. If there is a sufficient quality and quantity of unique detail to individualize (based on the ability of the examiner, including training, experience, etc.), then the examiner arrives at a conclusion of individualization. In essence, by disproving or rejecting the hypothesis of inconclusive, the null hypothesis has been overwhelmingly supported: individualization. If the thorough comparison had not revealed a sufficient quality and quantity of information to tip the scales, then the examiner would not have have disproved the hypothesis - that insufficient quality existed for an identification. Even though similarity may have been noted, the volume (or reliability) of that similarity would not have been sufficient to establish agreement necessary to prove their null hypothesis; therefore the examiner would have maintained the opinion/hypothesis of inconclusive that he/she began with.

In set B, if it can be disproved that sufficient dissimilarity does not exist to exclude, then there is sufficient dissimilarity (disagreement) and therefore exclusion. Again, by disproving or rejecting the hypothesis of inconclusive, the null hypothesis has been overwhelmingly supported: exclusion. If there had not been a sufficient volume or clarity of dissimilarity for exclusion, then the examiner would not have disproved the hypothesis. Even though dissimilarity may have been noted, the volume (or reliability) of that dissimilarity would not have been sufficient to establish disagreement necessary for exclusion; therefore the examiner would have maintained the opinion/hypothesis of inconclusive that he/she began with.

EVALUATION

22A) Hypothesis 6a: Sufficient Q/Q of similarity does not exist to establish agreement / individualization.
         6aHo: Sufficient Q/Q of similarity exists to establish agreement / individualization.
22B) Hypothesis 6b: Sufficient Q/Q of dissimilarity does not exist to establish difference / exclusion.
         6bHo: Sufficient Q/Q of dissimilarity exists to establish difference / exclusion.



Through this diagram it is easy to explain our conclusions to a jury.  If we cannot establish agreement to a sufficient level for individualization, we can still state that we did not exclude the impression because there was a lack of dissimilarity and the presence of similarity.  "Due to the presence of similarity, the latent print could not be excluded as being the right index fingerprint of John Doe."  Likewise, if we cannot establish difference to a sufficient level for exclusion, we can still state that we did not individualize because there was a lack of similarity and the presence of dissimilarity.  "Due to the presence of dissimilarity, the latent print could not be individualized as the right index fingerprint of John Doe."

There has been some excellent discussion on this topic on the CLPEX.com forum over the last week, and I would encourage those interested in the concept of inconclusive determinations and their subsequent reporting to visit the board and participate in or at least read the posts.

Verification is simply the action of all elements of ACE being objectively complete by another competent examiner:

VERIFICATION

23) Action 12: All elements of ACE are objectively repeated by another competent examiner

From the definitions of these four possible conclusions, we can begin to look at error during the entire latent print process and define what it means. But this will be the subject of next week's Detail!

Consider the following anonymous story related to me for publication on a similar matter several years ago:

***************
With 20 years of experience, I have made many identifications that other examiners would not verify. For many years I did not understand why the others would not sign, but after training & watching new examiners develop I now know that each examiner works at a different level of expertise & many examiners could not see my identifications.

Our unit has evolved & now has 5 senior & 3 junior examiners, who are now able to verify more of my “borderline” identifications.

Our policy is to require 2 examiners to sign a latent identification with a third examiner to sign as an administrative verification. If 2 examiners sign an ID yet no other examiner will sign as #3, we will send the comparison outside of our agency. If only one examiner will sign an ID, then it is considered “not of identification value” (NIV). That is hard for me to take sometimes because I feel if I had to testify on the case, how would I testify to the latent as a “NIV” or “no match” when I think it is an ID? This is a dilemma for me ethically.

*****************

The topic next week involves errors, and I would like to include some anonymous personal stories to go along with each one. I think this will make the topic come alive, so please submit your narratives by reply e-mail to KaseyWertheim@aol.com.

Has one of the following scenarios affected you in a big way?

1) Someone threw away a latent print you would have kept and identified

2) You missed an identification because you thought it was a palm print when it was a fingerprint (or vice-versa), or you got the orientation wrong in the finger itself (double-loop whorl)

3) you pushed the envelope too far on an identification you weren’t sure about… but the person you identified pled guilty, and nobody else knows you weren’t sure

4) you knew the prints matched, but someone else wasn’t so sure… so against your better judgement, you reported the ident as “inconclusive” but later found out the person went on to harm someone

5) there wasn’t quite enough of the known print present in order to positively exclude a subject as having been the donor, but your “no match” report was used by an attorney as an exclusionary report… and you went along with it because it was the easy way out

6) you found a smudged latent print that wasn’t sufficient for identification, but was definitely a pattern type that the suspect didn’t have. You were trained that if there isn’t enough for an identification, then it isn’t “of value” so you threw out the latent print and reported it as a “No Value” case even though you knew it wasn’t the suspect’s print.
 

_____________________________________________________________________
Feel free to pass The Detail along to other examiners.  This is a free newsletter FOR latent print examiners, BY latent print examiners. There are no copyrights on The Detail, and the website is open for all to visit.

If you have not yet signed up to receive the Weekly Detail in YOUR e-mail inbox, go ahead and join the list now so you don't miss out!  (To join this free e-mail newsletter, enter your name and e-mail address on the following page: http://www.clpex.com/Subscribe.htm  You will be sent a Confirmation e-mail... just click on the link in that e-mail, or paste it into an Internet Explorer address bar, and you are signed up!)  If you have problems receiving the Detail from a work e-mail address, there have been past issues with department e-mail filters considering the Detail as potential unsolicited e-mail.  Try subscribing from a home e-mail address or contact your IT department to allow e-mails from Topica.  Members may unsubscribe at any time.  If you have difficulties with the sign-up process or have been inadvertently removed from the list, e-mail me personally at kaseywertheim@aol.com and I will try to work things out.

Until next Monday morning, don't work too hard or too little.

Have a GREAT week!