ACE-V The only methodology?

[S. Siegel]

I have been reading “Speaking as an expert” by Stephen C. McKasson and Carol A. Richards (a very good book). In chapter 2, The Meaning of Identification, pg. 31, it states “A well established methodology, one that is accepted by your particular scientific community, can be defended by pointing to its acceptance.”

You can find several methodologies in the literature. McKasson and Richards use Examine, Compare, Evaluate and then Verify.

They state “The standard technique of trained observation are; examine the impression, compare the impression to a known impression. Therefore, the evaluation is the first procedure in which anything unusual might arise…….” They go on to explain how the whole process is cyclical and that each time you reach the end of the process you find yourself back at the beginning. Meaning you can examine in greater detail.

The methodology generally accepted by the fingerprint community is ACE-V. (How many agencies actually state in their SOP that they use ACE-V)

SWGFAST guidlines state ACE-V as;

A-Analysis-the assessment of a friction ridge impression to determine suitability for comparison.

C-Comparison is the direct or side-by-side observation of friction ridge detail to determine whether the detail in the two impressions is in agreement based upon similarity, sequence and spatial relationship.

E-Evaluation is the formulation of a conclusion based upon analysis and comparison of friction ridge impressions.

V-Verificationis the independent examination by another qualified examiner resulting in the same conclusion.

ACE-V starts with analyze. Analyze is defined as separating the whole into parts for individual study. We do analysis of fingerprints as a whole. Can analysis be a part of the study as well as the study? I have no problem with the comparison part, but then ACE-V states evaluate and then to verify.

I see analysis as the whole process and evaluate as the first part. Why analyze and evaluate in the same process or even as McKasson states examine and evaluate? Why say Evaluation? Why not say Conclusion?

In an article by Dave Grieve, “The identification process…,” published in the Feb. 1988 JFI he states;

E - Evaluation: The initial step is to evaluate the latent print for potential evidence. All impressions must be examined to determine what information they contain.

C – Comparison: The types of characteristics in the latent print and their relationship to each other are studied and compared to the known print. This is a series of individual decisions and judgments that lead to a conclusion.

F – Finding: When all data has been analyzed and interpreted, one of four conclusions must be reached. (Some sources use the word Opinion or Decision or Conclusion)

V – Verification: After a decision has been reached then it goes to another examiner for verification.

Another problem I have with ACE-V is there are only three conclusions that can be reached; 1) individualization 2) exclusion 3) inconclusive.

What is done with the latent images that do not meet these criteria, those that are non-suitable?

To follow ACE-V you would need to do a pre-evaluation and then apply ACE-V only when the latent images are suitable for comparison, would it look something like this; 1) Pre-Evaluation 2) Analysis 3) Comparison 4) Evaluation (again) 5) Verification (PACE-V)?

Dave Grieve proposed four conclusions; 1) it was made by that person 2) it was not made by that person 3) it can not be identified to any one person, that is, insufficient characteristic agreement to reach a conclusion, thus the latent is unsuitable 4) no conclusive finding can be reached due to insufficient inked standards.

I have always felt ECO-V (opinion in place of finding) was a better fit for the identification process. It is simple and something I feel I can explain to a jury.

Which brings up my next problem, when asked what my methodology is, how could I truthfully answer this question?

Q. The methodology accepted by the fingerprint community is ACE-V. Is that the methodology you use?

A. To answer truthfully I would have to say that I do not use ACE-V in its currently accepted SWGFAST format.

For several years I have struggled with ACE-V trying to get it to fit the comparison process. I was wondering if I was alone.

Sandy

[Pat A. Wertheim]

My conclusion after observing and teaching a number of latent print examiners from different countries who explain what they do in different ways, is that the human brain works the same no matter what country you are from or how you explain the process. I believe a lot of the discussion or debate results from semantics, not from actual differences in the way we go about the examination process. Charles Parker has referred to the "Wertheim model" as a five step process, but in the article I wrote, I simply tried to show that ACE-V fits into a more classical description of the scientific method. Michele Triplett describes "Hypothesis Testing" as the process we use, but again, I see Michele's description as merely another way to explain the same process others call ACE-V. As I said, I believe our brains all work the same. I believe the real struggle is in trying to articulate a complex mental process we cannot fully understand.

[David L. Grieve]

Sandy, SWGFAST presents ACE-V as "a" methodology, not the only methodology. The adoption of the methodology guideline occurred after Mitchell, and in the Mitchell hearing, David Ashbaugh explained methodology from his perspective, that is, the ACE-V model first proposed by Huber and expanded by Tuthill. Initially, Huber and Tuthill believed the model did not fit latent print examinations as well as it did other impression areas, but when Ashbaugh argued that it did, the model was accepted, at least by the RCMP. When Mitchell arose and the Daubert hearing was scheduled, the entire government team was in agreement they would follow the ACE-V method in testimony to maintain consistency and avoid confusion. I have no objection to that particular model, nor to yours. As Pat wisely noted, how the process is explained may differ from person to person, for, in truth, the process is extremely similar if not the same. I like very much Michelle's explanation of hypothesis testing, for whether we formally state a hypothesis or not, we are performing that type of test. I teach ACE-V is that element of the scientific method in which one both the hypotheses and the null hypothesis, either supporting or refuting one or the other. If neither can be supported, the result is inconclusive.

In the nearly 20 years since I wrote that article, my thinking has evolved. I still prefer finding to conclusion or opinion, but I include in the possible findings (conclusions or opinions) that inconclusive due to insufficient standards is too restrictive. I prefer now to state an inconclusive finding is the inability to individualize or eliminate for any reason. But the critical issue to me regardless of what model is utilized is that the process is never linear, never only forward. Using the ACE-V model, analysis is the observation of what individualizing information is available. In comparison, analysis is often revisited as one tries to comprehend the available information as to accuracy and reliability, both in the unknown and known. In determining what agreement and disagreement is present, reanalysis is an ongoing occurrence as one observes the effects of deposition pressure, any lateral movement, substrate characteristics, etc. Until a conclusion is reached, I believe the examiner is constantly repeating the two earlier stages.

I worked with Steve McKasson for many years and we conferred often. I agree with you that he and his sister wrote a very good book, and I always found my conversations with either one fascinating. I also agree completely with his premise that the key factor is general acceptance of the methodology. Certainly ACE-V is, by now, known, but I've heard some rather incomplete or confusing explanations of the process that indicate the individual is reciting someone else's concepts. I think it is often ineffective to use someone else's explanations without first acquiring personal ownership of those ideas. When something doesn't fit or seems awkward, then the concept must be tailored for comfort. It appears you have done just that and your reasoning is sound.

[Pat A. Wertheim]

When McKasson & Richards state, “A well established methodology, one that is accepted by your particular scientific community, can be defended by pointing to its acceptance,” I have to believe that was in a pre-Daubert world. One of the Daubert criteria is "general acceptance," but that alone would not likely succeed in defending against a Daubert motion to exclude fingerprint evidence.

I agree with Dave that not all those who try and explain ACE-V in court do so effectively. Whether you use ACE-V to explain the examination process or you use some other scheme, your effectiveness relies on your understanding of the process you are describing and your ability to articulate it.

So, whether you use ACE-V or any other explanation of the methodology you use, the key is understanding and describing the process. Our brains may work the same, but not all explanations are equal.

[Michele]

Whether or not our brains work the same is an interesting topic but regardless of the answer I don’t believe our thought process is our methodology. I think a methodology needs to be clearly articulated, clearly understood, and the thresholds, standards, or criteria need to be stated before anyone can say they use a methodology correctly. It’s these elements that change our conclusion from being subjective to being more impartial. Having standards and criteria also make it possible to test a conclusion to see if it’s up to the standard. Some descriptions of ACE-V seem to be differences in semantics but some explanations are clearly procedural differences.

For example, is A the assessment of a friction ridge impression to determine suitability for comparison (as stated in the SWGFAST QA guidelines) or is it the methodical examination of friction skin impressions; separation into parts so as to determine
the nature of the whole (as stated in the SWGFAST glossary)? Or is the first definition one of several results that can be derived from the second definition? Is E forming a conclusion about consistency, sufficiency or both? Is V reproducing the result, confirming the result, trying to disprove the previous conclusion, or peer review? If V isn’t peer review then where does peer review fall into our methodology? Do we need it for our methodology to be considered scientific? Is blind verification the same testing method that other sciences call blind testing? Is there a place in the ACE-V methodology for counting points or would this be considered a different methodology? Do you have to exclude with one dissimilarity? Do we have to go from the known to the unknown? What is our goal? Do we want correct conclusions, justifiable conclusions, or accepted conclusions? Does it matter? I would say it does matter because only one of these types can be upheld as a scientific answer.

Without clearly answering these questions how can anyone claim that it’s the same method we’ve always used and we’re all using it the same? How can we apply a method correctly without having it articulated to us? If a method isn’t articulated then I’d say that the method isn’t scientific but more of an intuitive process. I like equating ACE-V to hypothesis testing because that’s how Huber initially explained it and it’s accepted as a valid scientific method. If someone is changing it slightly (like assuming that reproducing the result is good enough and the rest of peer review can be forgotten about) then this revised version may not hold up under Daubert as a valid scientific method. Slightly varied forms of ACE-V may constitute as valid scientific methods but the further you deviate from hypothesis testing the more criticism your methodology may get. Another consideration is that every time the method is slightly changed (due to our subjective understanding of it) then the new version may have a different error rate. How can we produce an error rate if the method isn’t clearly stated and followed?

I’m not trying to sound like I’m attacking anyone’s views, these are just questions I ask myself every day and I struggled to find any answers until I equated ACE-V with hypothesis testing, then the answers are easily found.

To answer Sandy's question, our sop's do say we use ace-v but even more importantly our sop's articulate how our agency expects examiners to use ace-v (which is more descriptive than SWGFAST's definition).

[Pat A. Wertheim]

I have never been convinced that rigid training in a specific methodology is a guarantee that a competent examiner will never deviate from that methodology. For example, take an examiner who is trained that it takes a minimum of eight points to constitute an identification, and if there are eight or more points then it must be an identification. Even though this examiner is strictly trained to eight points, he/she will see latents with seven points that are, without doubt, correct identifications. And our examiner will see latents with eight points that don't "feel" right. Or take it to sixteen points. Evett & Williams proved that, in the 16 point standard, if an examiner "knew" it was a correct identification, he/she would find sixteen points, whether they actually existed or not.

You raise some challenging and crucial questions, Michele. But the human mind works in ways too complex to summarize in a set of scientific formulae. I would like to hear from some of our cognitive scientists in this thread, researchers who are trying to learn more about the mental processes that lead a fingerprint expert to a conclusion of identification or exclusion. What about it -- Is the mental process determined by the methodology we are taught, or is the methodology we describe merely an imperfect attempt to define the mental process we instinctively use?

[David L. Grieve]

Michele, as usual your questions are pertinent and they address issues that need to be answered as best we can. I realized I should have qualified my earlier response by stating that I agree with Pat as to the methodology used by competent examiners being very similar if not the same regardless of articulation. I think your point that articulation is essential to define the method is a good one, and I believe we are gradually finding that common ground. The strength of SWGFAST is that development and adoption of guidelines is by consensus, and that is also a weakness. Impartial reviews of SWGFAST guidelines have been critical of their vagueness, particularly in the areas you mentioned, and I think that is a valid concern. At the same time, achieving consensus on language within a diverse group requires patience and compromise. Not ideal, perhaps, but a hopeful beginning.

Pat, I think it is way too early to have any definitive insight into cognitive recognition, but Busey and Vanderkolk have made progress. I noticed in the abstract for their presentation in San Diego that the data gathered thus far reveals differences in eye movement and brain activity between trained examiners and laypeople when performing the same task. I am curious as to whether trained examiners exhibit the same physiological responses or if there is variation. If the eye movements of experts are similar, perhaps that is an indication that the same method is being employed.

Sandy, I neglected to state that the ACE-V method is in our procedures manual, although rather brief in description. However, a detailed description is provided in the training manual.

[John Vanderkolk]

Cognitive vision scientists are our friends. We need to study what they know, as we can be studied by them to better explain the visual system of light to eyes to brain to decisions. Stephen E. Palmer, Vision Science, Photons to Phenomenology, The MIT Press, 2002, page 377, Chapter 8, Representing Shape and Structure, and Chapter 8.2, Theories of Shape Reperestation, ISBN# 0-262-16183-4;

“The reader is forewarned that all of the theories we are about to consider are inadequate to capture the astonishing power, versatility, and subtlety of human shape perception. How people perceive shape is certainly among the most difficult problems in visual perception, so difficult that no satisfactory solution has yet been proposed.”

The vision scientists know we can perceive and understand what we see. Explaining the entire process is the challenge. There are adequate explanations. These explanations can be better, as all science strives for better explanations.

and
Chapter 9, Perceiving Function and Category,
pages 413-414 Four Components of Categorizaton:

"Given that one often categorizes objects into known, functinal classes, how might that goal be accomplished? There are a very large number of possibilities, but all of them require the following four basic components:

1. Object representations; the relevant characteristics of the to-be-categorized object must be perceived and represented within the visual system.
2. Category representations; each of the set of possible categories must be represented in memory in a way that is accessible to the visual system.
3. Comparison processes; There must be some way in which the object representation is matched or compared against possible category representations.
4. Decision processes; There must be some method for deciding, on the basis of results of comparison processes, to which category a given object belongs. "

Sounds like the cognitive visison scientists are striving to better explain how we do what we do. Palmer writes of catergorization into a class of objects, as they study vision and perception. We individualize, exclude, and fail to individualize or exclude. Or, categorize into or out of a class of one unique and persistent source. Sounds like 1 is Analyze the unknown image, 2 is Analyze the known image, 3 is Compare the two images, and 4 is Evaluate the process.

McKasson's terminology, Grieve's terminology and Wertheim's terminology are very similar to ACE. I do not see much difference in any of them. Understanding and explaining what we do and how we do it is better than rote memorization of a process that even the experts of vision science are striving to better explain. In a few years we may call ACE something better.

[S. Siegel]

Maybe my problem goes back a little bit farther than the methodology debate. I have trouble saying what we do is science. Maybe that is why I have trouble getting it to fit. I feel the process is more comparative analysis. Start with two items and compare them to see if they came from the same source. Is that too simple?

What does the customer want (the jury)? An explaination of how I reached my conclusion. Having a method is a way to articulate what we do to a jury. I know what I'm doing I just need to explain it to the jury in terms that work for both sides.

I’m not sure about hypothesis testing. Would that be “form an idea and then test it?” Would that be a form of confirmation bias? I believe it’s a hit or no hit, now I need to prove it.

To answer my own question, we do not have ACE-V in our SOP’S at this time. For those that have ACE-V in their SOP's. I would like to see how you have it written. You can e-mail it directly, I'm in the book.

Sandy

[Heidi Fraser]

No, Sandy, I don't think we are talking about confirmation bias here. Your hypothesis is not "I believe this is a hit" -- or at least, it shouldn't be, if you are practicing science.

The way I look at it, you are setting up a null hypothesis of "this latent and this inked standard were not made by the same finger." Given the number of fingers in the world that could have laid down a latent, that seems like a pretty safe bet! THEN, you set out to prove that hypothesis wrong, thus accepting the alternative hypothesis, which is that they WERE made by the same finger. This is how the scientific method works -- set out a null hypothesis, prove it wrong, accept the alternative hypothesis.

In my mind, if you start out with a really easily acceptable null hypothesis (these two prints did not come from the same source) and then prove it wrong (they did come from the same source), you have really accomplished something! That should be a conclusion you can really stand behind because you have demonstrated something against staggering odds.

Obviously, you must have a good reason to point to these two impressions and say "I have shown that the null hypothesis is incorrect and these two impressions were made by the same source," because what a huge statement that is! Luckily, we do have a good reason. It's called careful analysis and evaluation using the ACE-V method. And just to make really, really sure we've reached the correct conclusion, we have our work independently verified by another qualified scientist, who should be starting with the same null hypothesis. Yup, sounds a whole lot like science to me.

Heidi Fraser, MSc. Biology, Duke University

[S. Siegel]

If you believe its science then you would be correct. I find it very hard to believe the comparison process is science. We are not contributing any new information to the body of knowledge. We are repeating the same process that has been done numerous times over many, many years. This would make it Applied Science at the most.

How many people actually sit down and say “these two prints are not from the same source and I going to prove it wrong”.

Let the evidence speak for its self. I have no preconceived notions.

1) I have two prints and I compare them for similarities looking for discrepancies. If there is sufficient agreement without any unexplained discrepancies then I make a decision as to identity.

2) If there is no agreement in ridge detail or something I can not explain through pressure or distortion, then I must exclude.

3) If there is not enough information in either the latent print or the inked for comparison then it would be inconclusive.

The decision is made at the end after a thorough analysis.

No degree, but still a damn good examiner.

Respectfully submitted,
Sandy

[Heidi Fraser]

Hmmm.. It appears we are hitting upon a difference of semantics again (as we so often do on these boards). You're correct -- comparison sciences are Applied Sciences. But they are still science. You don't have to be discovering something new to be practicing science. What you are describing is the difference between Applied Science and Research Science.

Both are valid sciences. Both rely on scientific principles and application of the Scientific Method. I guess when people say "this isn't a science" what they really mean is "it's not a research science." I would have to agree that it's not when it's being applied (which isn't to say there isn't great research being done in the field) but it is nonetheless a valid science. There are many applied sciences being practiced in this world, and many projects being done that ride a fine line between Applied and Research-Based science as well. In my mind, the reason we DO research science is to develop sound scientific principles that can then be applied. Sorry if this sounds like a soapbox; I'm not trying to attack you, I just happen to think there is nothing wrong with applied science. People say it like it is a dirty word. Applied Science is still science.

True, there is a lot of research that still needs done in our field. Show me a scientific field where everything has already been figured out. I don't believe there is one. However, that does not lessen the validity of what we do.

To answer your question, I don't necessarily sit down and consciously think out my hypotheses every time I sit down to compare. I think it has become something of an automatic background process in my mind. But I can tell you that when I am comparing one latent to a stack of tenprint cards, there are a whole lot of prints that you pass by almost immediately as "nope, that's not it." To my way of thinking, you have validated your null hypothesis and moved on. It's when you see one that looks like it sorta, kinda, maybe could be an ID that you slow down and mentally go, "Whoa! THAT doesn't fit my null hypothesis (i.e. none of these match). Better take a closer look at this one." In my mind, that is when the real hypothesis-testing begins.

Just my $0.02, no better or worse than anyone else's.

[Charles Parker]

John, I want to thank you for your post. I for one have no problem with cognitive scientist determining how the five senses interact with the mind to acquire, sort, store, and retrieve information. I would encourage that kind of research.

The four steps by Palmer is good information. However I seem to articulate them in a different way.

Object Representation: The Evaluation of the image for its value or worth.
Category Representation: The Recognition of reliable ridge events and their unit relationship.
Comparison Process: The Comparison of that unit relationship with a second image.
Decision Process: The Conclusion reached upon apprasial of the ridge events. (Or Decision--Finding--Opinion) [something to indicate a final cognitive function].

When all four components are combined we have Friction Ridge Analysis. Analysis = The breaking down of the examination into its component parts.

But like what has been said. It is just semantics. What really matters is how it is explained to those outside the fingerprint community. After all isn't that one of the purposes of ACE-V. Easy to remember and articuate.

I see Palmer did not put Hypothesis in the components. I did not put it in mine either because I am an Analyst and not a Scientist so it is not required.

Thanks for the post and the information proposed by Palmer.