Mother-Daughter close non-match.

[Cindy Rennie]

Does anybody know where I can find an article in the JFI regarding the close non-match between mother and daughter? Both of the fingerprints are whorls, and the cores are quite close, but they came from two different people. I am not a member of the I.A.I. and do not have access to the journal on-line. The article is not in the November/December issue. Does anybody remember what issue it is in?

Cindy Rennie

[Pat]

I seem to remember from my training in courses at the FBI back in the mid 1980s that they had a mother and daughter whose fingerprint patterns were so close that they were next to each other in the old FBI Henry classification files. It was said that the 10-prints look so similar as to be indistuingishable except on very close examination at Level 2. I think they said they would sometimes give a card of one to a new employee to classify to see if they would catch the situation. I never worked at the FBI and only heard the story in passing at a class there. Perhaps one of the old timers from the Bureau would remember more.

[Ernie Hamm]

Not that recent, but here is a list of some articles dealing with heredity in fingerprints. The McCann article dealt with family fingerprints.

“A Family Fingerprint Project”, James S. McCann, Identification News, 25:5 (1975)
“Double Billing”, John E. Berry, Fingerprint Whorld, 5:19 (1980)
“Heredity in Fingerprints”, Gaye Shahan, Identification News, 20:4 (1970)

Not able to provide the articles, just the source. For any newbies (sort of), "Identification News" was the IAI's forerunner of "Journal of Forensic Identification".

[g.]

Cindy,

Are you sure you are not referring to the December 2011 issue of IDENTIFICATION CANADA? Vol. 34, No. 4, 2011?

It has a cover with a very nice 12 ridges in "relative agreement" close non-match from a mother/son pairing.

Does that help?

g.

[Bill Schade]

Glenn was kind enough to share this cover shot

photo.JPG

[Boyd Baumgartner]

Maybe Glenn could be kind enough to generate a likelihood ratio on this as well.

[g.]

With the FSS shut down, it is not possible to use their model at this time.

However, there is no need to. I can estimate having run a number of close non-matches through the models that this would generate a fairly high LR. Probably at least an
LR > 100,000 using the FSS model. But that should not be a surprise to anyone who understands how the models work. Since the current models only consider the level 1 and level 2 detail, this should create a fairly high LR due to the high degree of visual similarity for a higher number of minutiae. That also fits with our empirical observations. A high LR for this one makes sense because it is rare to have a large number of minutiae in agreement like this (especially not the result of an AFIS search). That is precisely why they put it on the cover of a journal. If it was so commonly observed, I predict the LR would be much lower (as it would undoubtedly be if you only had, say, 3 or 4 minutiae in relative agreement).

When the new model(s) is/are available it certainly can be run. But again, in this example, it is somewhat irrelevant since the L3D allows for the discrimination and would not be ignored by human analyst. A more interesting question posed by ID Canada is "what would happen if the images were of lower quality with minimal to no level 3 detail to discriminate"?

g.

[timbo]

Or maybe - and call me crazy on this, folks - just maybe, if the prints weren't so heavily cropped, they would be even easier to tell apart just on level 1 & 2 detail...... I'm guessing there are some very obvious dissimilarities just outside where the images were cropped - hence the reason why they were cropped so tightly......

[Bill Schade]

Well, while I don't disagree with you on your premise, that was not the point of the discussion

Since most latent impressions are partial, the issue would be is it concievable that prints with this much similarity in a small area might fool someone in casework

And since latents are rarely this clear, isn't this something to keep in mind as we make identifications?

[Neville]

But you would need your head examined to make an ID on 6 points! Particularly when they are out of alignment.

[kevin]

I'm along the same line of thinking as you Neville - not so much that I wouldn't make a call off of 6 points, but the fact that 4-5 of those are just such common (I call them 'passive' sometimes) minutiae formations for a loop - bifurcations above the core. I found myself looking for something else to anchor on and really didn't find anything. When searching loops I tend to notice the ones that don't have bifurcations, or a series of bifurcations, above the core.

With the research that is being done on rarity of minutiae formations I would think some of the new stat models could/would take that into account and maybe this print would get rejected as no value. Or maybe stat modeling does that already?

[ER]

.....rejected as no value.

Remember to specify 'no value for ID'. Either print would have plenty of 'value' for exclusion if that somehow became necessary.

[kevin]

Remember to specify 'no value for ID'. Either print would have plenty of 'value' for exclusion if that somehow became necessary.

Yeah - I meant for identification...

[Pat]

.....rejected as no value.

Remember to specify 'no value for ID'. Either print would have plenty of 'value' for exclusion if that somehow became necessary.

Remember to specify 'no value for ID'. Either print would have plenty of 'value' for exclusion if that somehow became necessary.

Yeah - I meant for identification...

SWGFAST wrestled with the issue of "of value" because we recognized that different agencies operate under different paradigms. When we wrote the document "Standards for Examining Friction Ridge Impressions and Resulting Conclusions," we addressed two approaches. One is the use of the criteria, "of value for identification." The other approach is "of value for comparison." Both approaches are valid and both are in widespread use, although an examiner working in an agency and an area where only one is used may not understand the validity of the other approach.

For a discussion of these two approaches, see section 5.1.4.2 of the document at
http://www.swgfast.org/documents/examin ... ns_2.0.pdf