I'm writing up a Phd dealing with ancient fingerprints. As it's an archaeology PHD and as archaeologists know nothing about fingerprints, I'm forced to include copius sections all about everything having to do with fingerprints and fingerprinting.
I'm looking for a reference that gives statistical proof of fingerprint uniqueness. Apparently there was an FBI AFIS study sometime around 2001 that said the chances of two different people with the same four points in comparison was 1 x 10 to the 27power! Does anyone know the details of this study and where I can get my hands on a copy? Thanks! Please e-mail any answers to k.moran@forensicoutreach.com
Statistics on Uniqueness
-
David Fairhurst
- Posts: 196
- Joined: Wed Jul 06, 2005 4:11 am
- Location: UK
- Contact:
Kimberlee,
You can never have statistical PROOF of anything. Only statistical SUPPORT. That is the nature of statistics, they will always give you some chance however small, even if the true figure is zero.
The 50k study you refer to supported the theory of uniqueness of fingerprints with a figures of 1 in 10 to the power of 97 against 2 whole fingerprints matching each other, and 1 in 10 to the power of 27 for fragments containing 15 minutiae matching more than one whole print.
It must be stressed that this study relied solely on minutiae placement on an AFIS system. This is only a very limited portion of the information available to the latent print examiner; but still the figures are well in excess of those that are relied upon day-by-day for DNA matches. I take this to be very good supporting statistical evidence of the theory of biological uniqueness.
I don't know where to get a copy of the report, maybe someone else will be able to help.
You can never have statistical PROOF of anything. Only statistical SUPPORT. That is the nature of statistics, they will always give you some chance however small, even if the true figure is zero.
The 50k study you refer to supported the theory of uniqueness of fingerprints with a figures of 1 in 10 to the power of 97 against 2 whole fingerprints matching each other, and 1 in 10 to the power of 27 for fragments containing 15 minutiae matching more than one whole print.
It must be stressed that this study relied solely on minutiae placement on an AFIS system. This is only a very limited portion of the information available to the latent print examiner; but still the figures are well in excess of those that are relied upon day-by-day for DNA matches. I take this to be very good supporting statistical evidence of the theory of biological uniqueness.
I don't know where to get a copy of the report, maybe someone else will be able to help.
-
Mike French
- Posts: 41
- Joined: Fri Jul 22, 2005 1:45 pm
- Location: Washington
- Contact:
Kimberlee,
You may try reading "Advances in Fingerprint Technology, Second Edition". There are summaries of 10 different statistical models including the first 50K study. All of them have inherent problems, as it is difficult to measure biological uniqueness with statistics. However, I believe as a whole they do support individualization.
I recently heard that the 50K study is being done again with new sets of data and new parameters that address problems with the first 50K study. I am looking forward to the results.
Regards,
Mike French
mike.french@metrokc.gov
You may try reading "Advances in Fingerprint Technology, Second Edition". There are summaries of 10 different statistical models including the first 50K study. All of them have inherent problems, as it is difficult to measure biological uniqueness with statistics. However, I believe as a whole they do support individualization.
I recently heard that the 50K study is being done again with new sets of data and new parameters that address problems with the first 50K study. I am looking forward to the results.
Regards,
Mike French
mike.french@metrokc.gov
-
Alan Christensen
50K study
From what I've gathered the FBI study has never actually been published, but here's a link to the testimony of Steve Meagher of the FBI latent print section on the 50K study. This is from the first of the Daubert fingerprint hearings.
http://m15080.kaivo.com/LegalDev/NLADA/ ... ll%203.PDF
http://m15080.kaivo.com/LegalDev/NLADA/ ... ll%203.PDF