>> And this guy who claimed in cyberspace to be a lawyer gave you stats you
>>
>> could check?
>
> I wonder if there is a means of gathering such stats? Is there a
website that keeps track of the kinds of evidence in court cases,
whether one person's word, DNA evidence, video evidence, etc?
*ALL* evidence comes down to taking the word of one or more people.
That includes DNA, fingerprints, footprints, trace blood samples,
tire tracks, bite marks, video evidence, bullets left in bodies,
guns found in the possession of the defendant, eyewitness testimony,
forged signatures on documents, etc. There's no such thing as
"self-proving" evidence. If you have a video of the defendant
committing the crime, you still need testimony that the crime happened
within the jurisdiction of the court, and within the lifetime of the
defendant. Also people can't be uniquely identified on videos.
And usually no one person provides testimony for enough elements
of the crime to convict alone. (e.g. one person testifies about
collecting DNA from an alleged rape victim, one person testifies
about collecting a reference sample from the defendant, and another
person testifies that the two samples match, plus the victim testifies
that the act was not voluntary but she can't identify the rapist
as he put a bag over her head from behind). That's *worse* than
convicting on one person's word. (There are some common exceptions:
domestic violence situations, alleged date rape, a cop pulling you
over for holding your cell phone while driving with no passengers,
etc.)
With DNA evidence, you have to take the word of:
- The person who collected the sample at the crime scene that this
is where the sample came from.
- The person who collected the reference sample, that this is who
the reference sample came from.
- The person who provided the reference sample (who might have used
fake ID to frame someone).
- The person(s) who processed the sample, that it was done correctly.
- The person(s) who maintain the lab equipment, that contamination did
not happen.
and further, if any *ONE* of those people is lying (or simply mistaken),
your conclusion is toast.
Video evidence requires testimony as to where and when the video
was taken, that timestamps on the video, if any, are accurate, and
that the video was not tampered with afterwards. The best case you
get with video is where someone in the video who survived the crime
testifies that he was there and this is what happened. In the worst
case you have to trust half a dozen people that the videos did not
get mislabelled or mixed up, and you have to trust *ALL* of them
to believe the video evidence. It would be very bad if the same
video could be used to convict someone robbing a local 7-11 last
month *AND* convict Lee Harvey Oswald of killing JFK, and we'll
assume that it's not a multi-year-long video.
Physical evidence like a body with an odd-shaped hole in the skull
and an odd-shaped weapon that matches the hole still needs testimony
that (a) connects the weapon to the defendant, and (b) the body was
killed as a result of *this* crime. Also, matching hole to weapon
may not be as obvious as looking at both. You also have to deal
with the possibility that there are thousands of similarly odd-shaped
weapons like this out there, and a different one was used in the
crime.