Are there any recent credible studies as to the likelihood of a person
being wrongly convicted of a felony given the burden of proof in the
United States and the various fallibilties, biases, type of crime,
population subgroup, representation quality, or part of US?  Likewise,
what are the chances that someone who actually comitted a crime would
be acquitted?  I would like to have credible studies, not speculation,
otherwise a good background on why such studies aren't being done. 
Essentially, what is the risk to a person and society of our standards
of proof and the legal system as a whole?  Respected speculation on
why no one is studying this would be acceptable is no one is worrying
about this issue.

I think lots of people and organizations are concerned about the
issue, but I can't see how any reliable statistics can be obtained.
Nearly everyone who is imprisoned claims to have been wrongly
convicted. Estimates of the actual number of wrongful convictions are
likely to be influences by political agendas.

In my comment above, I meant to type "influenced" rather than "influences."

The people who have the most to fear from our (or any) criminal
justice system are the innocent. Think about that for a minute.

If a guilty person is arrested and charged for a crime that (s)he
committed, the WORST thing that will happen is that person will be
punished for something they did. Anything less than that and they have
gotten away with something.

If an innocent person is arrested and charged with a crime, the BEST
outcome they can hope for is being acquitted from something they never
did. So after incurring expense, loss of freedom (and likely job,
family, respect), they are exactly where they started.

That's why we all need to fear a police force that is insufficiently regulated.

Good comments.  This an area that really bugs me, particularly with
the advent of the Innocence Project (I think they use DNA to rule out
or confirm death penalty cases - I know little else about it.)

A study of this question would probably require a very scientific and
careful study of the factors that cause guilty or innocent verdict. 
For example, I believe witness identification and other reliabity has
been studied, the fallibility of hair and fingerprint identification
is not being looked at and some other factors.  If someone had looked
at how these and other factors made up a verdict, that may give some
clues to the question.  Also, most defendants only get a few days of
legal time and a few more of defense, if they don't just plead guilty
due to the risk of being (possibly) wrongly convicted.  A scientic
study could spend months of research on a case without a predetermined
POV or agenda.  Thanks to both pinkfreud and markvmd for good points
here.  I hope I get an interesting and enlightening answer which
surely won't be definitive.