Slashdot Mirror
The Scientific Method Versus Scientific Evidence In the Courtroom
An anonymous reader writes "A few months back, the National Research Council and the Federal Judicial Center published the Third Edition of the Reference Manual on Scientific Evidence, the primary guide for federal judges in the United States trying to evaluate scientific evidence. One chapter in particular, 'How Science Works,' written by David Goodstein (Professor of Physics and Applied Physics at CalTech), has raised the issue of how judges should see science in the courtroom: should they look at science to see if it matches our idealized view of the scientific method, or should they consider the realities of science, where people advocate for their own theories far more than they question them?"
realities of science, where people advocate for their own theories far more than they question them
If you're in a courtroom, you should ALWAYS assume than anyone presenting evidence has an agenda (because they almost always do). No defense or prosecution attorney is going to put a scientist (or any other witness) on the stand who is going to do anything but advocate for their version of the case. Any judge who isn't completely new or blind already knows that well.
What political party do you join when you don't like Bible-thumpers *or* hippies?
Good watch: http://www.pbs.org/wgbh/pages/frontline/real-csi/
One thing they didn't cover, however, was the horror behind "expert" psychologists/psychiatrists and the damage they inflict.
If you're in a courtroom, you should ALWAYS assume than anyone presenting evidence has an agenda (because they almost always do).
I don't really agree with this. It's the lawyer's jobs to try to find a scientist somewhere who will vouch for their client or could provide evidence for the client. Now, ideally the scientist is from some other place that never heard of this case or knows anything about the case other than what he/she is an expert on. The science has been done long before hand, is sound, has been peer reviewed, etc, etc. The problem that I garnered from this article is when an "expert" is presented yet they are not peer reviewed, their science is not sound, nobody in their community takes them seriously, their degree is a hundred years old or questionable, etc, etc. Or they know a lot about the case and they have stepped forward to voluntarily promote their agenda to back their science by building credibility via high media courtroom cases.
The point is that when scientists come into a courtroom as expert witnesses (or really any expert witness), the only agenda on their mind should be to relate to the court what they have discovered in their research. Not how much more funding they'll get when this hits the papers. Not how much the defense is paying them to say that in their professional opinion the client is insane.
This is about the validity of science presented in trials, not whether or not the defense or prosecution as an agenda. I don't think that's ever been under question or they're a terrible defense/prosecution and the client should move for a mistrial. Psychological testimony has gotten so far out of hand that some states have taken extreme measures.
My work here is dung.
(At least) two different meanings of "science" are at work here. There's science as in "what scientists do," which is indeed a collection of methods and processes; and there's science as in "the body of relevant scientific knowledge," which is what juries need to know to decide on cases. Both are valid meanings, and it's important not to confuse the two.
The correlation between ignorance of statistics and using "correlation is not causation" as an argument is close to 1.
should they look at science to see if it matches our idealized view of the scientific method, or should they consider the realities of science
How about both? Look to see if the real process aligns closely with ideals, then base decisions on that. Did the lab follow best practices to prevent contamination? Were the statistics compiled by someone who knew what the samples were? Do all of the numbers include error calculations?
The discerning judge who considers scientific evidence will end up with a subjective opinion of whether the result meets the need for accuracy. It's the judge's job to apply such subjective opinions fairly, and here that means allowing only evidence that meets their realistic ideal.
You do not have a moral or legal right to do absolutely anything you want.
That the way real men do it!
Sent from my ASR33 using ASCII
Its a Hidden Markov Model problem. There is some underlying theory with a probability of being correct based on research. But the people presenting evidence in court do so through an additional set of weighting factors that govern their testimony based upon the underlying theoretical model. The testimony is what the judge and jury are able to observe. From it, they need to deduce the underlying model and probabilities.
Good luck expecting 12 people to understand this who have to have football (US) plays explained to them by TV announcers.
Have gnu, will travel.
No, we should do things following the method of a traditional witch trial: Throw the suspect in the lake. If they don't sink to the bottom and drown, they're a witch and should be burned at the stake without delay.
Alternately, you might also weigh the suspect against a duck ...
I am officially gone from