Point Standards

[Pat]

There is a difference between using a point standard and counting points. A "point standard" is a predetermined threshold below which no identification can be made. "Counting points" merely means that you have quantified the level 2 details present and corresponding between two fingerprint images. There are other ways to "measure" or "quantify" detail in a latent print, and it is true that research shows examiners do not always define points the same nor count them the same. Nonetheless, point counting is an attempt to quantify and, in the absence of trying to assign a minimum threshold prior to analysis, there is nothing wrong with knowing how many corresponding points are present between two prints after an identification has been made.

[Boyd Baumgartner]

Hi Lisa,
I've found a case where the 10th Circuit Court of Appeals affirmed a conviction in which the examiner stated they used "7 points of congruency" filed 7/18/2006.

The decision can be found here (it will want to download an rtf file) , but the relevant portion of the appeal and claim are quoted here:

During trial, the government’s expert witness, Barbara Crane, who performed the fingerprint identification for the Utah Bureau of Criminal Identification (“BCI”), testified that she had matched Mr. Cruz’s fingerprints with a number of critical deportation records. The training by the FBI requires at least ten points of congruency to establish a fingerprint identification match. The training that Ms. Crane received from BCI requires at least seven or eight points of congruency to make a fingerprint match. On one of the deportation records, the Warrant of Deportation of June 14, 2000, the fingerprint identification match was made with seven points of congruency. The fingerprint identification matches with regard to the Administrative Removal Order of September 10, 1999, and the Record of Sworn Statement from September 8, 2004, were made with eight points of congruency.

Mr. Cruz was found guilty and sentenced to 21 months in prison.

Mr. Cruz appeals the following two issues: (1) that the trial court erred when it changed the jury instructions to read, “defendant was deported at any time prior to September 8, 2004” rather than, “defendant was excluded from the United States on or about June 14, 2000;” and (2) the trial court erred in permitting the government’s fingerprint expert to testify that the fingerprints matched when fewer points of congruence were used than required by the FBI.

[L.J.Steele]

Thanks, Boyd.

The official cite if anyone else is interested seems to be United States v. Cruz, 189 Fed.Appx. 725 (10th Cir. 07/18/2006). It is an unreported decision, so there may be restrictions on how it is cited and used.

The decision says:
"During trial, the government's expert witness, Barbara Crane, who performed the fingerprint identification for the Utah Bureau of Criminal Identification ("BCI"), testified that she had matched Mr. Cruz's fingerprints with a number of critical deportation records. The training by the FBI requires at least ten points of congruency to establish a fingerprint identification match. The training that Ms. Crane received from BCI requires at least seven or eight points of congruency to make a fingerprint match. On one of the deportation records, the Warrant of Deportation of June 14, 2000, the fingerprint identification match was made with seven points of congruency. The fingerprint identification matches with regard to the Administrative Removal Order of September 10, 1999, and the Record of Sworn Statement from September 8, 2004, were made with eight points of congruency."

Then
"Ms. Crane stated on direct examination that she has been doing fingerprint identification for 24 years, has compared 3,336,000 separate impressions, and has never incorrectly identified a fingerprint match. During trial, the critical exhibits 15-C and 4-A, showing the Warrant of Deportation of June 14, 2000, and Fingerprint Card of September 17, 2004, displayed the fingerprint identification being made with seven points of congruence. However, Ms. Crane stated that there were three more points of congruence that were not listed on the chart, making in total, ten points of congruence, meeting the FBI standard. In light of Ms. Cranes qualifications and her testimony that there were at least ten points of congruence, the admission of her testimony was not plain error."

This sort of thing worries me. Expert says I charted 7. When pressed with a need for 10, the expert says on the stand, there's 3 more that I didn't chart. (No indication whether these were specified on the stand.) So the challenge is moot.

If there's a standard, then it should be in the documentation at the time.

[Boyd Baumgartner]

Another conviction affirmed on appeal.

STATE OF TENNESSEE v. LEON FLANNEL
Direct Appeal from the Criminal Court for Shelby County
No. 03-07354 Carolyn Wade Blackett, Judge
No. W2007-00678-CCA-R3-CD - Filed October 13, 2008


http://statecasefiles.justia.com.s3.ama ... eonOPN.pdf

Page 4

Nathan Gathright testified that he worked with the Memphis Police Department as an expert
in latent fingerprint examination and identification. According to Mr. Gathright, a fingerprint found
on the cardboard beer container matched the defendant’s fingerprint at seven points. Mr. Gathright
noted that it was widely accepted that a seven point match of a fingerprint constituted a positive
identification.

[Michele]

Pat,

You make an important point that, as experts, we don’t all count points in the same way. This makes the number even more meaningless than I originally thought it was.

I can see that ‘point counting is an attempt to quantify’ but in my opinion it’s a poor attempt since it doesn’t quantify anything. Instead, it may mislead both layman (detectives, prosecutors, defense attorneys, and judges) and fingerprint practitioners into believing a conclusion has more or less weight than it does. Does an ID with 20 points indicate it is beyond debate (to quote Locard) while an ID with 7 points imply that it’s a borderline conclusion? All examiners should be able to easily see that this isn’t true and be able to explain why. Giving a number without other information is as misleading as stating that conclusions are absolute and conclusive.

there is nothing wrong with knowing how many corresponding points are present between two prints after an identification has been made

I can’t agree. Noting such information implies to me that points have more weight than other elements (like rarity, clarity, intervening ridges, incipient ridges and dissimilarities). It also implies that more is better. If someone feels the need to quantify the weight of a comparison, it seems to me that they should note all vital elements, not just one.

With the advent of statistical models on the horizon, noting a point count could imply that a likelihood ratio is needed for lower count ID’s but not needed for higher count ID’s. Is this true? If it were true then some of the most notable errors would have been considered so definitive that the need for a statistical model might not have been deemed necessary. From my own research on errors, most errors are due to examiners misinterpreting the characteristics in a print and not due to the same amount of similarity in another person. This would lead me to believe that objectivity (clarity of an image) is just as important as the interpretation that a point exists. And of course, the rarity of the features are important too.

[Pat]

. . . as experts, we don’t all count points in the same way. This makes the number even more meaningless than I originally thought it was.

Because we don't all count points the same way, the number of points we come up with may vary from one examiner to the next, or even with the same examiner from one day to the next. But that doesn't render the number totally meaningless. If an examiner tells me an identification has 22 clear points, I can reasonably believe that identification has more detail in it than if the examiner were to tell me in the next case the identification had only 6 clear points. I agree with you that the number of points alone does not tell the whole story, but the number is not "meaningless."

______________________________________________________________________________________

there is nothing wrong with knowing how many corresponding points are present between two prints after an identification has been made

I can’t agree. Noting such information implies to me that points have more weight than other elements (like rarity, clarity, intervening ridges, incipient ridges and dissimilarities). It also implies that more is better. If someone feels the need to quantify the weight of a comparison, it seems to me that they should note all vital elements, not just on.

I differ with some of the other teachers on this issue. While I agree that clarity, incipients, etc., all carry weight, I also maintain that points are still the backbone of what we look at. And I readily accept that there are exceptions, such as the "Zero Point Ident" published in the JFI a few years ago. You can pull out exceptions all day long to prove that a point count is not always a reliable indicator, but in the vast majority of cases a point count still gives us information of value to some degree in the analysis of the latent print, or in an identification.

Too many people have made the leap that because we don't use a minimum number of points as a threshold before we even see the latent, it is wrong to count points under any circumstances. A predetermined threshold is wrong for the very reasons you cite. But to extend that reasoning to deny that a number has any value at all is also wrong.

[L.J.Steele]

Does an ID with 20 points indicate it is beyond debate (to quote Locard) while an ID with 7 points imply that it’s a borderline conclusion? All examiners should be able to easily see that this isn’t true and be able to explain why.

If an examiner tells me an identification has 22 clear points, I can reasonably believe that identification has more detail in it than if the examiner were to tell me in the next case the identification had only 6 clear points. I agree with you that the number of points alone does not tell the whole story, but the number is not "meaningless."
[snip] I also maintain that points are still the backbone of what we look at. [snip] You can pull out exceptions all day long to prove that a point count is not always a reliable indicator, but in the vast majority of cases a point count still gives us information of value to some degree in the analysis of the latent print, or in an identification.

From an attorney PoV, I'm mostly likely to get a call from a defense attorney thinking about a print challenge if there's a single ID with low number of points. Prosecutors and defense attorneys will likely make different closing arguments, in good faith, about that 22 point ID vs the 6 pt one. The Appellate Courts likewise may be more receptive to arguments about the relevance and admissibility of the 6 pt one vs. the 22 point one. It matters to a lay audience -- and it may help to figure out ways to explain to the bench, bar, and public, when and why it matters to you folks.

[Michele]

I'm not denying the value of points but there is only value to the points when considered in conjunction with other factors (especially clarity). If a defense attorney calls me about a case that had a 22 point conclusion, I wouldn't have faith in the conclusion; I'd want to know how clear those points were. If there are 22 CLEAR points, and in sequence, I'd agree, that indicates weight. If the points are UNCLEAR then the weight is significantly diminished. The number alone tells me nothing. My personal opinion is that it is dangerous for practitioners and attorneys to give weight to a conclusion based solely on the number of points. If we continue to ignore the importance of other factors then I doubt that we'll improve the accuracy of our results.

Just thinking about this, it seems like some factors are less important when the number of points is very high. Rarity, for example, means a lot when the number of points is low but doesn't mean so much when the number of points is high. Clarity, on the other hand, always seems important to me because I've seen errors with very high point counts listed but the clarity was extremely poor.

The examples supporting my view may be rare but these are the specific cases we're trying to account for. It's these cases, erroneous conclusions, that we're trying to protect against, aren't we?

[ER]

The problem in the above two cases is that the experts are telling the courts that the number of points is the critical thing, and that once you reach that number, then the ID is made.

Once the experts start telling the court that an identification is more complex than 'points', and that the entirety of the print is compared (including the stuff between the points), then the court may begin to understand more of the nuances of our discipline.

[Neville]

Hi Michele
I am glad I held back from saying what I thought, based on what I thought you were saying. I now must agree with you. Back in the dim dark days of the 12 point rule, we operated under an agreement with Australia that 12 clear points were required for positive identifications. However then it became a disagreement about what was clear and what was not, the word clear had not been defined. Shirley's so called fingerprint identification in Scotland is a clear case of what you are talking about if there is a set number of points (17) and that is all you need; then you do get into the situation of teasing the points this never a good place to be, I can always imagine what is going on in a print but to prove it is something else again. There is always a case for counting points for your own benefit but then as my friend John would counter why count if you know it is him, and that is hard to argue against.

[C. Coppock]

There are some great points here. (Pun intended)

We should not report LV 2 counts of characteristics/points/minutiae or count them in an "effort to quantify" our comparison, because a simple number count does not offer insight to the complexity of the comparison.

With a simple example of LV2 only: 6 minutiae can have 15 unique spatial relationships. However, a different set of six with distortion blocking areas of the print will have less. Example should minutiae 1,2,3,4 not have a reasonable line of sight to minutia #6 for an “accurate” analysis, this would represent a 40% reduction in available spatial relationships. While some of these gaps may be covered with an estimate of the spatial relationships, it invariably lowers the value of the information. Accordingly, these are not equivalent information sets at level 2 even if both sets contain 6 each.

[sandra wiese]

Michele, perhaps I can shed some light on this from a whole other perspective...tho likely just add to the mud in the water. ; )

In discussions with other examiners, I have encountered some who record the number of matching points in their notes and or count them before testimony (really) specifically to head off this line of questioning from defense attorneys. For quite awhile locally (and I do know there are differences to lines of questioning in different areas of this country and the rest of the world) one repeated line of defense questioning has been in regard to the number of points used in an ident. Apparently this speaks to their whole "fp science is too subjective to be allowed as expert testimony" idea. So it is my understanding that some people DO do this so they can just give a number and move on to the other questions.

No, it isn't me so I can't tell you why they think this is good and when I ask regarding following this line of reasoning to its completely illogical conclusion, I can't get a straight answer. I have found this LARGELY to be with people doing ident only cases with ten print to ten print comparisons and not so much with the latent to ten print comparisons. However, the personnel doing this is split between latent examiners and ten print examiners.

This line of "reasoning" is also why these same people will do a ten print ident (for a habitual offender or possession of a weapon by a previous offender case, both of which are very common locally) and initial the whole CARD as opposed to just one exemplar space to head off the question of "why did you only compare one finger on the card?"

Again, I have zero issue with getting asked these questions in court and I don't record my points and do only compare one finger on a ten print ident, so I cannot tell you exactly what is going on. But my feeling is that these things are done more to lessen the amount of defense questioning they get as opposed to any valid, scientific reasoning. Court scares the pee out of a LOT of people, as you know, and some will go to great lengths and twisted logic to avoid or reduce the experience. It isn't a matter of not understanding the science but rather a tremendous fear of public speaking at play.

[sandra wiese]

oops, and just a quick drawing out six points and connecting the dots with not much effort, I have already found twenty unique spatial relationships. I'm of course using three points per relationship, which in my head is a requirement to define a unique point in space. Am I missing something? just curious from an intellectual exercise point of view, not a "oh oh I can have TWENTY points with only 6 points" point of view! haha

[Shane Turnidge]

To Lisa's point, I think it is a reasonable assumption for any counsel to question the validity of a friction ridge identification based in part on 6 or 7 points.

We as examiners know there is more to the ACE-V process than counting points but it is a question that is commonly asked by cousel on both sides of the equation. When I'm asked "how many points do you have?" I answer the question. Then I go one to talk about how L2D is a slice of the spectrum of information I use to arrive at my conclusions and I go on to articulate the remainder of the information.

From an attorney perspective, when confronted with low point count friction ridge identifications, I would be asking how was the clarity in both the latent and exemplar. For more information on the subject consider the sufficiency graph in SWGFAST's draft on Standards for Examining Friction Ridge Impressions and Resulting Conclusions (Ver 1.1)
I'd also be asking how much of the minutiae were affected by pattern force in their development vs. those minutiae that were truly random. And I'd be asking to see the examiners contemporaneous notes on their analysis of the latent print. When examiners test their thresholds they had best be prepared to defend their positions.

Shane Turnidge

[C. Coppock]

Regarding the unique relationships of minutiae, I was referring to level 2 relationships that are not counted more than once. In our simple case of 6 minutiae, a relationship between points #1,#3 would not be counted again from a perspective of #3,#1. This would simply be one relationship.
The attached document I have run the numbers out, counting all relationships only once. Of course this represents an ideal situation where all points are within a simple linear perspective of each other. In many cases the numbers would be less due to intervening distortion.
Since all the relationships are relevant to an individualization, I considered them as a set.