Well this article from the LA Times seems to be the talk of the criminal defense community. Read the article here.
This entry was posted
on Wednesday, July 23rd, 2008 at 10:47 pm and is filed under In The News.
You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
5 Responses to “A discovery leads to questions about whether the odds of people sharing genetic profiles are sometimes higher than portrayed.”
The silence is deafening….but I shall talk to the brick wall for a moment.
Another HUGE problem that seems to be overlooked by many is when the DNA analyst has the profile of the defendant PRIOR to even doing their testing of whatever evidence they are given to test. When they have the defendants profile and they are confronted with some “grey” area’s that are subject to interpretation, it makes it quite easy for them to make the call that fits the defendants profile.
If they were not allowed to have access to the defendants profile PRIOR to coming up with the profile of the alleged suspect or a deduced profile, there would be no potential bias in the resulting interpretation.
An interesting and accurate write up of this problem can be found here:
So if you take Troyer’s test/study and this article about
sequential unmasking together, is DNA evidence, when it comes to identifying people,
anything other than junk science?
Do our brothers/sisters in Forensic Science know of any motions
drafted that cite these studies? So that DNA evidence might be excluded on basis of
unreliability?
I don’t interpret either articles as coming anywhere near claiming that DNA is “junk science”. They are really two somewhat separate issues…both of which need exposing.
1) You have a statistical model that may not be what in fact it has been held out to be.
2) You have lab analysts who are using your clients profile as a “cheat sheet” to come up with the “right answer.” (the answer that gets your client convicted)
If you have a case involving DNA, refer it to FSD. There is no need to “wing it” on your own..
Supplemental discovery requests should be made in DNA cases asking for any FBI memoranda to the labs discussing reliability of or challenges to statistical models, actual searches that contradict current statistical models, or actual searches that show matches at 9 or more loci. Recent CODIS database searches requested by defense attorneys may reveal that the actual probabilities may be closer to1 out-of-100’s instead of 1 out-of-trillions or billions, which makes a huge difference. Numerous actual matches have been found at 9 to 13 loci! This means that the statistical models put forth by prosecutors may simply be wrong. In fact, they are probably wrong because their statistical models are based on a random match between unrelated individuals. However, the evolutionary science used in the first place for DNA suggests that all individuals are, in fact, to a degree, related. So the assumption used by the State is flawed to begin with, as born out by actual matches, rather than the flawed theory used by the prosecutors. What this means, ultimately, is that where there is not a match, your client should be excluded. Where there is a “match,” it is just simply another piece of circumstantial evidence of guilt, rather than actual evidence of guilt. Furthermore, an alleged match should be examined because your client may, in fact, not match at several loci where judgement calls must be made to determine whether there is a match or not. This is where having the target’s DNA may influence the examiner to make a call one way instead of another at a questionable loci. So, two issues: First, current statistical models are flawed because CODIS internal matches and evolutionary science of relatedness both contradict prosecution statistical theories. This is probably Brady material and internal memoranda evidencing this should be tendered to the defense. Second, DNA evidence is most probably actual evidence of innocence, but only circumstantial evidence of guilt.
I do agree with Chris. DNA is powerful and clear as to exclusion. As to inclusion, based on the recent news, it is much less indicative of guilt. It is more than a possibility but far less than virtual certainty which is where the prior statistics left us.
July 26th, 2008
The silence is deafening….but I shall talk to the brick wall for a moment.
Another HUGE problem that seems to be overlooked by many is when the DNA analyst has the profile of the defendant PRIOR to even doing their testing of whatever evidence they are given to test. When they have the defendants profile and they are confronted with some “grey” area’s that are subject to interpretation, it makes it quite easy for them to make the call that fits the defendants profile.
If they were not allowed to have access to the defendants profile PRIOR to coming up with the profile of the alleged suspect or a deduced profile, there would be no potential bias in the resulting interpretation.
An interesting and accurate write up of this problem can be found here:
http://www.bioforensics.com/articles/sequential_unmasking.html
July 28th, 2008
So if you take Troyer’s test/study and this article about
sequential unmasking together, is DNA evidence, when it comes to identifying people,
anything other than junk science?
Do our brothers/sisters in Forensic Science know of any motions
drafted that cite these studies? So that DNA evidence might be excluded on basis of
unreliability?
July 28th, 2008
I don’t interpret either articles as coming anywhere near claiming that DNA is “junk science”. They are really two somewhat separate issues…both of which need exposing.
1) You have a statistical model that may not be what in fact it has been held out to be.
2) You have lab analysts who are using your clients profile as a “cheat sheet” to come up with the “right answer.” (the answer that gets your client convicted)
If you have a case involving DNA, refer it to FSD. There is no need to “wing it” on your own..
August 4th, 2008
Supplemental discovery requests should be made in DNA cases asking for any FBI memoranda to the labs discussing reliability of or challenges to statistical models, actual searches that contradict current statistical models, or actual searches that show matches at 9 or more loci. Recent CODIS database searches requested by defense attorneys may reveal that the actual probabilities may be closer to1 out-of-100’s instead of 1 out-of-trillions or billions, which makes a huge difference. Numerous actual matches have been found at 9 to 13 loci! This means that the statistical models put forth by prosecutors may simply be wrong. In fact, they are probably wrong because their statistical models are based on a random match between unrelated individuals. However, the evolutionary science used in the first place for DNA suggests that all individuals are, in fact, to a degree, related. So the assumption used by the State is flawed to begin with, as born out by actual matches, rather than the flawed theory used by the prosecutors. What this means, ultimately, is that where there is not a match, your client should be excluded. Where there is a “match,” it is just simply another piece of circumstantial evidence of guilt, rather than actual evidence of guilt. Furthermore, an alleged match should be examined because your client may, in fact, not match at several loci where judgement calls must be made to determine whether there is a match or not. This is where having the target’s DNA may influence the examiner to make a call one way instead of another at a questionable loci. So, two issues: First, current statistical models are flawed because CODIS internal matches and evolutionary science of relatedness both contradict prosecution statistical theories. This is probably Brady material and internal memoranda evidencing this should be tendered to the defense. Second, DNA evidence is most probably actual evidence of innocence, but only circumstantial evidence of guilt.
August 5th, 2008
I do agree with Chris. DNA is powerful and clear as to exclusion. As to inclusion, based on the recent news, it is much less indicative of guilt. It is more than a possibility but far less than virtual certainty which is where the prior statistics left us.