[from Jeff Vancouver 940801.11:12]
to Avery, Bill L. and Martin (numerous posts)
Thanks for the response. I got what I wanted. Bill L. your zero
reference signal is simply a version of my second option (higher-order
sends a generic reference signal). The exact nature will depend on the
level/unit one is refering to. It this point I am interested in keeping
it general.
Avery gave me a third possibility (P becomes R). Of course, there needs
to be a mechanism that regulates that connection. But that idea that
three possibilities exist works for me for now.
Still no response to my free will posts?
Paul George 940729 10:00
The issue of selection is complex (that is why we have a field devoted to
it). Getting the job description is done via job analysis. A central
problem you did not mention is that the jobs often change (as will be the
case with interacting systems, I argue). But the issue of Type II error
and compensating skills is basically one of averages and individuals. The
individualist can always argue that someone will get messed over via these
tests (for some will). The averagist will argue that, on average, more
are helped than hurt (which is also true). The point I am trying to make
is that selection will occur, thus some individuals will get messed over,
the best we can do (at this point) is to try to minimize the number of
people. Thus we continue to strive to make better instruments. To say we
cannot help everyone so should not try to help anyone is morally
irresponsibility. (I am not going to attempt to help that droning person
because there is a chance I might not be able to save her and that she
might save herself is I just let her be).
The irony of the polygraph situation is that I witness testimony (the only
non-circumstantial evidence) is even more unreliable than a polygraph.
The best answer to these dilemmas is to develop collabrating evidence.
Each single piece of evidence is just more evidence for making the case.
The higher level issue of everyone using the same test and thus some never
being selected is one that I worry about, particular since we have found
how useful and global general cognitive tests are. So far, this has not
been a problem supposedly, but I think a clever sociologist needs to look
into it.
Later
Jeff