I was watching a Nova program today titled, "Forensics on Trial".
The documentary was marginally decent, and overly dramatic and overstepping in its bounds regarding many subject matters.
However, one matter that I found rather interesting was what wasn't mentioned once during the show. In fact, the thought caused me to reflect back on all the years of reading in casual interest on forensic analysis because I couldn't recall a single time the matter had ever been addressed.
So far as I'm aware, forensic analysis is only required to show that evidence does match by an interpretation of the measurements, and some measurements require more interpretation than others.
These measurements are produced with a probability attached, and this probability is produced (most often) by a standard degree of probability for something to look like a match and yet not be a match.
The documentary highlighted the recent call by Congress to review forensic analysis to rate how much science was actually in the forensic science field.
The only form of evidence that really walked out without a beating was DNA evidence.
All other forms of evidence were virtually torn apart and received very poor ratings in their reviews.
What surprised me was that when I thought back, I couldn't recall a single instance where I have ever read a forensic case which showed a null hypothesis against an hypothesis.
There's simply a compare and contrast and the best match is selected if enough of a match is made in whatever form of evidence is being analyzed (not counting DNA).
The documentary interviewed a few folks involved at various upper levels within the industry who stated that the Congressional review has effectively shaken the trust out of current forensic analysis until the field can tighten back all evidence analysis methods to a degree comparable to DNA.
I don't see that happening any time soon, so to me it seems as though this kind of goal setting really causes a massive issue in the near-future.
But why not just apply a basic scientific principle to this field that is supposed to be scientific?
Why not just apply the null hypothesis method?
In fact, since we have one lawyer attempting to prosecute and one to defend; why isn't their a forensic team for the hypothesis and a team for the null hypothesis?
Currently, the only means of a pseudo-null hypothesis that I'm aware of is at request of the defense team using a counter expert to review the evidence.
However, by this point, the material being reviewed is after-the-fact, and not at the same time and within the same publication as the original forensic analysis.
It would seem to me to be good science, one wouldn't think of conducting a true/false conclusion of an hypothesis without running it against a null hypothesis and supplying openly the deviation from the material evidence both contain; thereby showing a degree of probability in the supplied evidence by scientific standards.
-Just a random thought after watching a documentary on the holes left in current forensic analysis.
The documentary was marginally decent, and overly dramatic and overstepping in its bounds regarding many subject matters.
However, one matter that I found rather interesting was what wasn't mentioned once during the show. In fact, the thought caused me to reflect back on all the years of reading in casual interest on forensic analysis because I couldn't recall a single time the matter had ever been addressed.
So far as I'm aware, forensic analysis is only required to show that evidence does match by an interpretation of the measurements, and some measurements require more interpretation than others.
These measurements are produced with a probability attached, and this probability is produced (most often) by a standard degree of probability for something to look like a match and yet not be a match.
The documentary highlighted the recent call by Congress to review forensic analysis to rate how much science was actually in the forensic science field.
The only form of evidence that really walked out without a beating was DNA evidence.
All other forms of evidence were virtually torn apart and received very poor ratings in their reviews.
What surprised me was that when I thought back, I couldn't recall a single instance where I have ever read a forensic case which showed a null hypothesis against an hypothesis.
There's simply a compare and contrast and the best match is selected if enough of a match is made in whatever form of evidence is being analyzed (not counting DNA).
The documentary interviewed a few folks involved at various upper levels within the industry who stated that the Congressional review has effectively shaken the trust out of current forensic analysis until the field can tighten back all evidence analysis methods to a degree comparable to DNA.
I don't see that happening any time soon, so to me it seems as though this kind of goal setting really causes a massive issue in the near-future.
But why not just apply a basic scientific principle to this field that is supposed to be scientific?
Why not just apply the null hypothesis method?
In fact, since we have one lawyer attempting to prosecute and one to defend; why isn't their a forensic team for the hypothesis and a team for the null hypothesis?
Currently, the only means of a pseudo-null hypothesis that I'm aware of is at request of the defense team using a counter expert to review the evidence.
However, by this point, the material being reviewed is after-the-fact, and not at the same time and within the same publication as the original forensic analysis.
It would seem to me to be good science, one wouldn't think of conducting a true/false conclusion of an hypothesis without running it against a null hypothesis and supplying openly the deviation from the material evidence both contain; thereby showing a degree of probability in the supplied evidence by scientific standards.
-Just a random thought after watching a documentary on the holes left in current forensic analysis.
via JREF Forum http://forums.randi.org/showthread.php?t=263664&goto=newpost
Aucun commentaire:
Enregistrer un commentaire