UPDATES ON CLPEX.com

TENPRINT STANDING COMMITTEE

The Scientific Working Group on Friction Ridge Analysis, Study and Technology (SWGFAST) was formed to establish consensus guidelines and standards for the forensic examination of friction ridge impressions. To date, that has been limited to the concerns of the latent print community. Although it has been realized that there are areas of common interest to other fingerprint applications (e.g., criminal history and biometrics), existing SWGFAST guidelines were not developed with the intention of being applied to non-latent print related matters.

Individuals from the IAI and FBI CJIS, having realized a need to provide similar guidelines for the Tenprint community, contacted SWGFAST for assistance. At their request, consideration was given to having SWGFAST expand its role to include this aspect of the profession. This resulted in the decision to implement a Standing Tenprint Committee within SWGFAST.

To address this need, individuals whose primary responsibility is in performing tenprint examinations and have positions requiring knowledge of end-to-end tenprint operations will be added to the SWGFAST membership. Individuals possessing these qualifications that are interested in serving on the Standing Tenprint Committee must submit their letter of interest to SWGFAST Executive Secretary, Maggie Black (mab@fss.ocgov.com). The letter of interest is to be included as an attachment, in Word format, to the cover email.

Each letter must include your qualifications, experience, professional affiliations, and what you bring to the group. Within the letter, or as an attachment, you must include confirmation of your agency’s support for you to actively participate in this working group, along with their support for you to attend the meetings as scheduled (SWGFAST normally holds two one week meetings per year). Please note that the FBI sponsors SWGFAST and covers all permissible travel, lodging and meal expenses related to these meetings. They are prohibited from paying the expenses of other federal agency employees and for the travel cost of international members. Include in the email subject line and as a reference topic on the Letter of Interest: “SWGFAST Tenprint Committee Membership Application - 2007”. All submissions should be received no later than August 31, 2007. The selection process is expected to be completed by late September.

Questions can be directed to SWGFAST chairperson, Leonard Butt, at lbutt@mdsp.org.

Best Regards,
Leonard Butt
Chairperson

_________________________________________

Last Week we had a clear-cut winner for the 2007 CLPEX.com T-shirt Slogan contest:

Registered Board Users: 491
Newsletter Subscribers: 1,973
Forum Posts: 3,684
Monthly Hits: 352,027

Convicting more bad guys
because of information sharing...
PRICELESS.


(CLPEX.com in a Mastercard logo configuration)

We continued a series on fingerprint 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.

We continue the series with a discussion of errors in the process.

_________________________________________
Errors in Latent Print Examination - Part 1 of 2
by Kasey Wertheim

Last week, we concluded a discussion identifying 23 required examination process milestones for an accurate conclusion:
 

23 REQUIRED EXAMINATION PROCESS MILESTONES FOR AN ACCURATE CONCLUSION

Factors (F), Actions (A), and Hypothesis sets (Hs) consisting of Hypotheses (H1) and Counter Hypotheses (H0) of Fingerprint Examination
 

PRE-ACE

Necessary Factors for Known and Unknown Impressions

1)       F1: Source was present

2)       F2: Matrix was present

3)       F3: Surface was present

4)       F4: Touch of source with matrix transferred to surface

5)       F5: Environment did not completely obliterate the
       resulting impression

Necessary Actions for Recovery of Unknown Impressions

6)       A1: A crime occurs and is investigated

7)       A2: Item at the scene is noticed

8)       A3: Item is documented and preserved for collection

9)       A4: Item is collected

10)   A5: Item is preserved for examination

11)   A6: Impression is developed or otherwise visualized

12)   A7: Impression is preserved, documented, or otherwise
       retained for analysis

ANALYSIS

Necessary Decisions (bold) & Actions necessary for Comparison of Unknown and Known impressions

13) Hs1: Impression is suitable for retention as potential case evidence

14) Hs2: Impression is suitable (quality and quantity of detail) for comparison (Tolerance is set)

15)  A8: Impression is correctly claimed (in terms of location) and oriented

16)  A9: Unique detail is viewed and analyzed

17)  A10: Relevant distortions are observed and accounted for; (Tolerance is adjusted)

18)  A11: A distinct group of detail is correctly targeted for comparison

COMPARISON

Necessary Decisions (in bold) for Evaluation of Unknown and Known Impressions

19)  Hs3: Sufficient common area is present in both the known and unknown impressions

20)   Hs4: Sufficient combined quality is present in both the known and unknown impressions for Evaluation

21)   Hs5a: No similarity is present between the unknown and known print
(H0: Similarity is present between the unknown and known print)

Hs5b: No dissimilarity is present between the unknown and known print

(H0: Dissimilarity is present between the unknown and known print)

EVALUATION

Necessary Decisions for Individualization or Exclusion of Impressions

22)   Hs6a: Sufficient Q/Q of similarity (agreement) of detail does not exist to establish individualization (INC.)

(H0: agreement exists to establish individualization)

Hs6b: Sufficient Q/Q of dissimilarity (disagreement) of detail does not exist to establish exclusion (INC.)
(H0: disagreement exists to establish exclusion)

VERIFICATION

Necessary Actions for Completion of the Accepted ACE-V Methodology

23)   A12: All elements of ACE are repeated by another competent, unbiased examiner

On the CLPEX.com Discussion Board this week, an examiner relates each action and hypothesis to a decision.  In fact, at 20 points in the process there is opportunity for a decision that could be either correct or incorrect.  In almost every instance, a bad decision can cause a correct conclusion to be missed - whether that conclusion was exclusion or individualization. 

If the first responder doesn't recognize the need for evidence, and in fact nobody recognizes the need for evidence, then no opportunity for collection would ever occur.  Likewise, if they recognize the need for evidence but don't actually preserve it for collection, that decision might lead to destruction of the latent print.  Yet another issue is that the evidence could be recognized and preserved for collection, but not actually collected.  Each of these errors in the decision making process results in the chance that a latent print would never be developed and therefore would never have the opportunity to be identified.

After the evidence is collected it must be preserved for development.  Incorrect packaging and storing an item so long that ridge detail deteriorates are probably among the biggest issues in today's crime lab environment.  With backlogs and staff shortages, the time to process incoming cases is a challenge with much less time to spend training officers to collect evidence correctly.   These and other factors can render a once-suitable latent print unsuitable for a conclusion when the impression is finally developed.

For every latent print / surface combination there is an ideal series and order of processes that must occur to maximize the quality of the developed ridge detail.  The goal of latent print processing is to increase the contrast between the ridges and the background so that enough ridge detail can be seen to reach a solid conclusion as to identity or non-identity.  The farther away from the "target" of the ideal process sequence, the less likely an examiner is to develop suitable impressions on the item.  In some cases this won't matter but in other cases it will.  Unfortunately, there is no way to know ahead of time whether skipping a processing step (or two or three) will affect a particular latent print.  Generally, examiners try to maximize the development process, but administrative concerns such as time, money, and the priority of other casework sometimes affect how closely an examiner operates to the ideal.  It has to be realized that administrative concerns will always exist, and the job of the examiner includes operating within agency policies and procedures.  Ideal SOP's and policies will guide the employee down the right path when making processing decisions that might affect whether developed latent prints will be able to lead to a conclusion.  If it is discovered that a conclusion was missed because of the lack of application of a particular technique or sequence, then fault lies in one of two places: the agency for not having adequate policies and procedures, or the employee for not following them.

After development, someone generally looks at the impression to determine suitability for retention as case evidence.  If the impression is not properly preserved, it cannot be analyzed for suitability.  Take for example a print developed with Iodine fumes and a technician that doesn't recognize the effect that even a short period of time has on the impression.  A suitable print might fade to unsuitability before it can be captured for examination by an expert.  Many times, the technician is also the person who preserves and examines the impression, so preservation of the impression for examination is not generally an issue.

The next decision point is whether or not to retain the impression as case evidence.  This is the first point at which an analyst observes unique detail within a latent print and makes a determination of whether or not suitability exists to retain the impression.  Generally, this analysis is conducted while the print is still on the item, or while the impression is still being developed.  Consider the consequences of error at this point in the latent process.  For this example of error, let's further define the scenario.  Let's first say that in fact the impression was suitable to identify the suspect, but the analyst incorrectly decided it was not suitable to retain as case evidence.  In this case, the analyst could be a first responder or other non-expert crime scene person dusting an immovable object.  On the other end of the spectrum the analyst might be a fully qualified CLPE in a laboratory environment.  Reasons for the bad decision may range from inadequate training to inadequate conditions (lighting) to inappropriate consideration of administrative time concerns (rushing), to failing eyesight, or even downright carelessness.  Although the root causes of not claiming a suitable impression as case evidence are varied, the end result is unmistakable - a missed opportunity for the identification of a subject.

We must also look at the opposite scenario - that in fact the impression was not suitable for identification by anyone.  In this case, the error would be for the analyst to retain the impression as evidence in the case anyway.  It is generally accepted in the discipline that there is far less negative impact to a case to retain an unsuitable impression than there is to not retain one that is suitable.  Therefore, most analysts err on the side of caution and retain all impressions that may be suitable for comparison, even if there is doubt.  The real-world impact of retaining an unsuitable impression is that someone else may waste time looking at it again.  The real world impact of not retaining a suitable impression could range from no effect (such as if the print belonged to the victim) all the way to the loss of life (if the crucial murder case ident is missed and the subject goes on to kill again).

Exactly the same factors apply to the next process point - whether to conduct a side-by-side comparison with a particular impression.  If the examiner determines that there is so little quality and quantity of ridge detail in a print that even a side-by-side comparison to an ideal exemplar would yield no conclusive determination, then the examiner has deemed the print not suitable for comparison.  In the situation where the impression actually was sufficient for identification, the real-world impact of this error in judgment could lead to a missed identification and potential loss of life.  Likewise, if the print actually was not suitable for comparison, but the examiner conducted a side-by-side comparison anyway, the real-world impact would be wasted time.

Any experienced examiner will recall times where the incorrect orientation of the print caused a missed identification.  Usually, the scenario involves other case identifications and there is not a large consequence to this error.  But to study this example, lets take the worst case scenario of a small latent print with very few ridge flow clues to assist the examiner in a determination of impression orientation.  These are the latent prints we really hate to see come our way because usually they are time killers.  You know the ones I'm talking about - it could be a fingerprint or a palm print in any area of the palm and it could be rotated in one of many different directions.  And in order to exclude even one person as having been the donor of the print, it requires hours of detailed comparison time.  If we use this example to consider the action of orienting the print correctly, we begin to see that in fact there is a hypothesis that goes along with this action.  During the later phase of comparison, you constantly ask yourself whether you have considered all possible orientations.  For a latent that does in fact match, if you decide that you have accounted for all orientations and you are wrong, you have just missed the identification.  For a non-match, there is no real-world consequence to cutting off the comparison early.  But of course we don't know ahead of time whether a print matches or not, so generally we push aside consideration of the probability this wasn't him (and therefore the risk/reward of missing the match / saving time) and press onward toward a complete comparison.  The complete comparison involves coming back to this point in the analysis and answering the hypothesis that you had the print oriented correctly at that location.  For a non-match, only after this hypothesis has been answered for every possible location is the examiner able to confidently proclaim an exclusion.  And of course if the print did in fact match, at some point in this process the examiner would have the print in the correct orientation at the correct location and the latent process would proceed to the next decision point - recognition of unique ridge formations.

Even an examiner who has arrived at this point - a correctly orientated latent print - may not recognize or consider unique ridge formations that are present.  Consider the value placed on unique ridge shapes around a scar, disassociated ridge detail, or even crease formations in the palm.  If an examiner is looking at prints that actually match but they don't consider that type of ridge detail unique, they may not even consider the impression suitable for comparison, much less make an identification that someone else would easily make.  The critical question in this scenario involves the root cause of the error.  For this action, it is easier to consider the overlap into comparison: In not recognizing unique ridge formations, did that examiner make a non-match decision that was technically incorrect?  Yes - in fact, the prints did match and the examiner did not arrive at that conclusion.  But was the examiner wrong in making that decision?  Sometimes it helps to look at the opposite perspective.  If an examiner didn't see something he considered unique, would you want him to identify the print?  Would you want someone to say you were wrong because you didn't consider something you felt was not unique?  Few would argue yes to this perspective.  In fact, it becomes apparent that this is a training issue, not a bad decision issue. 

If an examiner is not appropriately recognizing the uniqueness of particular friction ridge features, perhaps he should learn why in fact those features are unique in the first place.  Then he could assign the correct weight to them.  There are many proactive individuals in our discipline that take initiative to pursue this type of training on their own.  On the other hand, there are other individuals who sit back and wait for the department to take responsibility for their employees and arrange the training.  Still yet, there are other examiners who don't feel there is validity in these "other" types of detail, and therefore they choose not to pursue (or actively resist) training necessary to accurately make the match determination in this scenario.  There are even employees who are quite content doing things the way they always have, even though deep down they know they need to pursue training that leads to change.  But regardless of the circumstances surrounding the scenario, the root cause of this missed identification was failure of the examiner to consider the uniqueness of the available ridge formations.  Whether that failure was acceptable is an administrative question for each examiner or agency.

Even if every unique ridge detail is recognized, the other major issue in latent print examination is distortion.  During the analysis phase, the examiner assesses the nature of the distortion and considers the extent to which it has affected the appearance of features.  This process of accounting for distortion allows the examiner to compare prints which don't exactly match.  And obviously that is the nature of our discipline - no two impressions will ever exactly match.  Distortion becomes an issue when it is so subtle that the examiner doesn't recognize it as such, and instead considers it a dissimilarity attributable to a different friction ridge source.  In the Analysis phase, this would result in incorrect features considered for a target group, and ultimately may lead to the real-world worst-case consequence of not finding those features, missing the critical case match, and allowing the subject to continue their serial murder spree.  As with recognizing uniqueness, this action also carries over into the comparison phase - accounting for distortion during the side-by-side comparison.  This is perhaps the biggest contributing factor to erroneous exclusions that exists in our discipline today.  This is mostly due to combining the recognition of distortion with the one dissimilarity doctrine.  An examiner falls into this quagmire when she strictly adheres to the one dissimilarity doctrine and feels that one distortion artifact is a reliable ridge feature.  More specifically, in the scenario where the prints actually match but the distortion was severe, the examiner would incorrectly conclude that an element of the impression was a feature instead of an artifact.  In this scenario, the pertinent distortion was not recognized as distortion.  Is this the fault of the examiner?  We find ourselves in the same predicament as with uniqueness.  To state that the examiner should have been able to attribute the element as being a distortion artifact instead of a feature would be an administrative decision based on the root cause of the issue.  Perhaps the examiner has not been adequately trained to recognize or tolerate distortion.  If the examiner can not assign the correct weight to dissimilar appearance, perhaps she should learn how distortion can affect ridge formations in the first place.  Again, many proactive individuals and agencies encourage or pursue this type of training, but others don't see the value until it is too late and they are dealing with conflict surrounding erroneous exclusions.  The root cause of the missed identification in this scenario was that the examiner failed to consider the extent of distortion on the impression and instead classified the element as a feature rather than an artifact.  Whether that failure was acceptable is an administrative question for the examiner or agency.

In both of these scenarios, we have proceeded from the assumption that the prints matched.  However, there is also the scenario in both actions (uniqueness and distortion) where the prints actually did not match.  In these cases, errors in considering uniqueness or assigning distortion as a features would have no real-world consequences because the prints did not match anyway.  Considering non-matching impressions does not lessen the fact that an error occurred, but it does remove the real-world impact of the error.

The last error during the analysis phase of ACE-V involves incorrectly choosing and memorizing a target group of details.  There are many facets of selecting and searching a target group as described previously in this series.  If an examiner correctly proceeds all the way through the process but incorrectly chooses a target group, the result could be that during comparison the target group is not recognized and therefore the identification is missed with the same real-world consequences as before.  For non-matches, there is no real world impact. 

As we enter the comparison phase of ACE-V, our first decision point is whether the known print contains suitable area for comparison with the latent print.  For example, if an examiner looks at a latent print and determines that it is from the side of the finger, and also noticed that the officer didn't even roll the fingerprints on the inked print or live scan exemplars, he could conclude that the area is not present and therefore not even conduct side-by-side comparisons.  In our match scenario, if he were wrong and the area was in fact there, he would end up with our real-world worst-case loss of life result from a missed identification.  If the examiner were correct and the area was outside the represented area of the exemplars, he should report an inconclusive determination and request complete prints. 

Generally known print quality is not so bad that the examiner decides not to even look for the target group.  But we have all seen AFIS search results of a candidate impression that was literally so bad that we couldn't even see detail.  This is the decision point we are at - to decide that the known print doesn't even contain suitable quality ridge detail for comparison with our memorized target group.  If the examiner decides that the exemplar does not contain sufficient quality for comparison but was incorrect, then there actually was sufficient clarity for the match.  Never having compared the area, she would end up with our worse-case missed ident.  And like other scenarios, if the prints did not match she would still technically be wrong, but there would be no real-world consequence.  But it brings up an interesting administrative point - perhaps the corrective action should be different for an examiner who made a bad decision based on area or clarity resulting in a missed identification, but who didn't actually conduct a comparison of the two areas of friction ridge impression.

After determining suitability of the area and quality of the known print exemplars, we proceed under the hypothesis that there is neither similarity nor dissimilarity and we try to prove this through a comparison of the same area of two impressions bearing sufficient quality for comparison.  As we look for our target group in different fingers, we confirm that we see dissimilarity in one finger and proceed to the next finger.  The process of seeing dissimilarity affords us the ability to effect an exclusion of that finger.  As we discuss this process, realize that we are literally dwelling on a very small snap-shot in time that actually may only consume very brief amounts of comparison time, depending on the time required to find the presumed location (area) of the target group on the known exemplar.  If the area is immediately at the core or delta, this time may literally be a split second, followed by only a split second required to determine that sufficient dissimilarity with our target group is present that we decide to move to the next print.

If this decision is in error, then actually that area of the previous print in fact DID contain our target group of detail.  We simply didn't recognize similarity, or mistook something else such as distortion for dissimilarity and moved on to the next print.  The real-world consequence depends on whether the examiner considers that they have conducted an exhaustive comparison of the latent print.  Many examiners search several different target groups to insure they don't miss target groups that appear distorted and therefore out of tolerance with their expected appearance.  This practice actually affords the examiner several opportunities to recognize similarity and make the identification.  From detailed discussions and practice of this process, efficient examiners will often proceed from a very narrow, specific targeting practice early in the search of a latent print to a more open, thorough targeting practice to conclude the search.  In other words, the first search through a stack of known prints for a target group might be based on less tolerance for different-appearing detail, and less tolerance for that group to be located away from the presumed location within the finger.  By lowering tolerance, the examiner is able to operate in an extremely efficient manner, but is more likely to miss the target group if it appears slightly different.  Good examiners who correctly implement this process can zero in on the correct finger in an extremely efficient way compared to examiners who go straight into exhaustive comparisons on each and every finger.  The tradeoff is that they sometimes have to search several target groups several times in order to be exhaustive and therefore conclusive.

It is important to consider the range of comparison techniques that exist because it directly impacts the discussion of error during the comparison phase.  An examiner who on three occasions incorrectly determines that similarity of the target group did not exist, but conducts a fourth, exhaustive comparison and finds the target group and makes the match, in essence made no error upon the conclusion of the process.  Yet technically, the examiner made three erroneous decisions for the sake of speeding through targeting on their way to the final conclusion.  The net effect is that an efficient examiner will actually spend far less time over the course of 100 cases than an examiner who does not target with this technique.  And in fact, the net accuracy of final conclusions may be the same or even higher than the methodical examiner who thoroughly searches each finger to exhaustion before proceeding to the next finger.  However, discussion of each sub-portion of the technique must be considered within the context of the over-all technique of the examiner.

Having clarified targeting techniques, let's proceed with the knowledge that the targeting process is iterative and requires enough cycles to arrive at a correct final conclusion.  At each target comparison the hypothesis is that no similarity or dissimilarity is present, and based on what is seen by the examiner at that location, one of two hypothesis sets is used.  If dissimilarity is seen, the examiner then makes the hypothesis that sufficient dissimilarity does not exist to establish an exclusion to the target group.  This is generally followed up with further consideration of the detail within that area to collect further visual data that either supports or refutes the new hypothesis.  If the area is clear, this entire process of searching for the target takes only split seconds.  The examiner finds additional similarity or dissimilarity and the focus now shifts to the evaluation phase of ACE-V as the examiner considers whether sufficient dissimilarity has been found in that finger during the entire process in order to effect a final decision of exclusion.  If several target groups have been searched, the examiner weighs the quality and quantity of similarity or dissimilarity to arrive at a final conclusion of agreement or disagreement.  If both the latent print and the known exemplar are clear in the relevant areas, several target groups may not be necessary to effect a final conclusion.  If one or the other print is unclear, several iterations through several different target groups in several areas may be required before a conclusion is reached.  For each finger of a subject, this process applies.  In total, the process applied to all 10 fingers in the case of a latent fingerprint establishes the exclusion or identification of the individual as having deposited the latent print.

Next week we will be able to jump right into the heart of decision making in the evaluation phase of the ACE-V process through the detailed exploration of 8 scenarios that result from evaluation.  This final exercise will consider error for all possibilities of two ground truth states: matching and non-matching prints.  We will work from the inconclusive hypotheses and the conclusive null hypotheses we use to evaluate whether or not prints match, and we will look at specific scenarios that relate to each of the 8 examples.  This is probably the most interesting and difficult subject matter to consider, so consider Starbucks next Monday morning on your way to work, and we'll have some fun!