While that is fancy, I'd rather see people doing the legwork to be able to state with scientific certainty that fingerprints actually are a reliable means of identification. Right now they're only even admissible because of inertia.
Inertia? I am confused by this comment. Could you please explain?
Fingerprints (friction ridges) are actually formed at about week 14-17 of gestation. Since they are partly formed by the effects of varying pressures of fluid in the womb, not even identical twins have exactly matching fingerprints. It is impossible to duplicate the these pressures, thus no two fingerprints are identical. Same goes for feet- and toeprints.
Outside of the more individualized areas of the pads of the fingertips, there are still recognizable differences, even down to where the sweat pores are located on the friction ridge detail. Fingerprint experts estimate that there are approximately 125 comparable points in each square centimeter of friction ridge skin.
Identification through fingerprints therefore also becomes a matter of statistics (Yes, yes: Lies, damn lies, and statistics. Still.); the likelihood of two fingerprints having the EXACT same 12 or more points matching becomes 40 thousand zillion to one (I exaggerate for effect), much like DNA indentifications. If one indentifying characteristic is sufficiently proved to be unique, a court can accept the identification based on that one point match. (That's very much a long shot, though.)
There are 3 major types of fingerprints -- Loops (~65% of all prints are loops), whorls (~30%), and arches (~5%). There are some subcategories (tented arches for instance), but loops, whorls, and arches are the main ones.
Fingerprints experts claim all this. None of this can really be measured in any objective way - fingerprint analysis is only presented as self-proclaimed experts claiming that the fingerprints match, which is a subjective and unreliable measure. Fingerprint analysis comes down to whether the expert says the mark looks like the sample taken from the suspect, with no measure of uncertainty, and frequent disagreements on the same sample between different experts, or even the same expert approached after a couple years with with samples they've previously identified with absolute certainty.
As well, none of it has ever been verified by any sort of study. You'd think that before claiming that no two people have matching sets of fingerprints, there might have been some kind of science undertaken to try to verify this - taking clean samples of a statistically significant portion of the population, proving that the samples are unique, and then proving that these unique fingerprints then leave identifiable marks that can be traced back to the unique fingerprints. Something foundational to suggest fingerprint study even has merit, even before you start getting into the reliability of experts? It hasn't. These sorts of studies have never even been attempted.
In the US, the standard that forensic analysis has to meet to be admittable is called the Daubert standard. You can read the specifics of it here (http://en.wikipedia.org/wiki/Daubert_test), but of the five factors, fingerprint analysis - along with every other forensic technique except nuclear DNA analysis - can only even present information to support #5, general acceptance. But this standard was codified in the 1990s, and forensic industries - fingerprinting, ballistics, odontology, hair analysis, etc - have been around for a lot longer than that. If actually scrutinized, they wouldn't have a leg to stand on, simply because they've gotten by for decades or centuries on the basis of "Everybody knows fingerprints are unique" rather than even the most preliminary data to actually support that. But no judge wants to be the judge that says fingerprinting is inadmissible due to fundamental unreliability.
The National Academy of Science did a massive study of forensic techniques and analysis a couple years ago (http://www8.nationalacademies.org/onpinews/newsitem.aspx?RECORDID=12589). Short version, except for nuclear DNA analysis, it's all conjecture at best and intentional fraud at worst. "with the exception of nuclear DNA analysis, the report says, no forensic method has been rigorously shown able to consistently, and with a high degree of certainty, demonstrate a connection between evidence and a specific individual or source."
"In contrast, for many other forensic disciplines -- such as fingerprint and toolmark analysis -- no studies have been conducted of large populations to determine how many sources might share the same or similar features." "Studies should accumulate data on how much a person's fingerprints vary from impression to impression, as well as the degree to which fingerprints vary across a population. With this kind of research, examiners could begin to attach confidence limits to conclusions about whether a print is linked to a particular person." It's not that the data is unconvincing, it's that there is seriously no data whatsoever. Fingerprinting is only admissible because it's always been admissible, since before we expected people to show their work.
Certainly, I have never seen or read a study of extant databases of fingerprints definitively proving the premise that each fingerprint is unique. It sounds like a fantastic doctoral thesis project. I doubt you could get a large enough population to agree to give you their fingerprints for a new study (maybe try colleges or the jails/prisons). I certainly would not agree to participate in it. But I agree that such a study would add to the validity of the field, even if only to confirm the premises already held.
DNA cannot individuate between identical twins; fingerprints can. Just sayin'.
Fingerprint analysis was the first forensic science to formally institute professional certification. And a way to de-certify professionals (which they probably do not do often enough). I don't know how often or even if they are required to re-test to maintain their certification, but I remember my instructors saying they had to re-cert for their various jobs as forensics experts or crime scene folks.
Experts make mistakes. All the time, in every field. Not everyone graduates at the top of their class and has the best work ethic and eagle-eyes. And truly, fingerprint comparison IS more art than science. You learn the rules of whatever classification system you're using (there are at least three big ones-most English-speaking countries use the Henry system) and then you look at a zillion prints, classify them, and have someone more experienced check your work to make sure you got it right. Experience and expertise really, really count.
Even if the computer gives you potential matches for a fingerprint, a human being must review each potential match and either eliminate or confirm it. The technology is not sophisticated enough to do that (despite what TV shows would have you believe). And if you ask three experts about something, you'll get five opinions. I agree, the error rate could be vastly improved. But there will always be differences in how well a person does a task. There are reasons that you want "the best" guy.
I am not sure if any other forensic science has really been developed since the Daubert standard was initiated (1993) except for DNA analysis. It's the only one that has had to prove itself. Fingerprints have been in use for over a century. Precedent counts in the court system. The previous standard in most jurisdictions was Frye, which is still in use in some places, and calls for just general acceptance. Science is always changing or revising or clarifying previous results; forensics is required to make a decision based on what you have, right then, right there.
It's a flawed system, but it's still better than anything else.
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
(no subject)
Date: 2011-05-18 04:07 am (UTC)(no subject)
Date: 2011-05-18 04:24 am (UTC)(no subject)
Date: 2011-05-18 05:34 pm (UTC)Fingerprints (friction ridges) are actually formed at about week 14-17 of gestation. Since they are partly formed by the effects of varying pressures of fluid in the womb, not even identical twins have exactly matching fingerprints. It is impossible to duplicate the these pressures, thus no two fingerprints are identical. Same goes for feet- and toeprints.
Outside of the more individualized areas of the pads of the fingertips, there are still recognizable differences, even down to where the sweat pores are located on the friction ridge detail. Fingerprint experts estimate that there are approximately 125 comparable points in each square centimeter of friction ridge skin.
Identification through fingerprints therefore also becomes a matter of statistics (Yes, yes: Lies, damn lies, and statistics. Still.); the likelihood of two fingerprints having the EXACT same 12 or more points matching becomes 40 thousand zillion to one (I exaggerate for effect), much like DNA indentifications. If one indentifying characteristic is sufficiently proved to be unique, a court can accept the identification based on that one point match. (That's very much a long shot, though.)
There are 3 major types of fingerprints -- Loops (~65% of all prints are loops), whorls (~30%), and arches (~5%). There are some subcategories (tented arches for instance), but loops, whorls, and arches are the main ones.
(no subject)
Date: 2011-05-18 06:10 pm (UTC)As well, none of it has ever been verified by any sort of study. You'd think that before claiming that no two people have matching sets of fingerprints, there might have been some kind of science undertaken to try to verify this - taking clean samples of a statistically significant portion of the population, proving that the samples are unique, and then proving that these unique fingerprints then leave identifiable marks that can be traced back to the unique fingerprints. Something foundational to suggest fingerprint study even has merit, even before you start getting into the reliability of experts? It hasn't. These sorts of studies have never even been attempted.
In the US, the standard that forensic analysis has to meet to be admittable is called the Daubert standard. You can read the specifics of it here (http://en.wikipedia.org/wiki/Daubert_test), but of the five factors, fingerprint analysis - along with every other forensic technique except nuclear DNA analysis - can only even present information to support #5, general acceptance. But this standard was codified in the 1990s, and forensic industries - fingerprinting, ballistics, odontology, hair analysis, etc - have been around for a lot longer than that. If actually scrutinized, they wouldn't have a leg to stand on, simply because they've gotten by for decades or centuries on the basis of "Everybody knows fingerprints are unique" rather than even the most preliminary data to actually support that. But no judge wants to be the judge that says fingerprinting is inadmissible due to fundamental unreliability.
The National Academy of Science did a massive study of forensic techniques and analysis a couple years ago (http://www8.nationalacademies.org/onpinews/newsitem.aspx?RECORDID=12589). Short version, except for nuclear DNA analysis, it's all conjecture at best and intentional fraud at worst. "with the exception of nuclear DNA analysis, the report says, no forensic method has been rigorously shown able to consistently, and with a high degree of certainty, demonstrate a connection between evidence and a specific individual or source."
"In contrast, for many other forensic disciplines -- such as fingerprint and toolmark analysis -- no studies have been conducted of large populations to determine how many sources might share the same or similar features." "Studies should accumulate data on how much a person's fingerprints vary from impression to impression, as well as the degree to which fingerprints vary across a population. With this kind of research, examiners could begin to attach confidence limits to conclusions about whether a print is linked to a particular person." It's not that the data is unconvincing, it's that there is seriously no data whatsoever. Fingerprinting is only admissible because it's always been admissible, since before we expected people to show their work.
(no subject)
Date: 2011-05-18 10:20 pm (UTC)DNA cannot individuate between identical twins; fingerprints can. Just sayin'.
Fingerprint analysis was the first forensic science to formally institute professional certification. And a way to de-certify professionals (which they probably do not do often enough). I don't know how often or even if they are required to re-test to maintain their certification, but I remember my instructors saying they had to re-cert for their various jobs as forensics experts or crime scene folks.
Experts make mistakes. All the time, in every field. Not everyone graduates at the top of their class and has the best work ethic and eagle-eyes. And truly, fingerprint comparison IS more art than science. You learn the rules of whatever classification system you're using (there are at least three big ones-most English-speaking countries use the Henry system) and then you look at a zillion prints, classify them, and have someone more experienced check your work to make sure you got it right. Experience and expertise really, really count.
Even if the computer gives you potential matches for a fingerprint, a human being must review each potential match and either eliminate or confirm it. The technology is not sophisticated enough to do that (despite what TV shows would have you believe). And if you ask three experts about something, you'll get five opinions. I agree, the error rate could be vastly improved. But there will always be differences in how well a person does a task. There are reasons that you want "the best" guy.
I am not sure if any other forensic science has really been developed since the Daubert standard was initiated (1993) except for DNA analysis. It's the only one that has had to prove itself. Fingerprints have been in use for over a century. Precedent counts in the court system. The previous standard in most jurisdictions was Frye, which is still in use in some places, and calls for just general acceptance. Science is always changing or revising or clarifying previous results; forensics is required to make a decision based on what you have, right then, right there.
It's a flawed system, but it's still better than anything else.