What is a "Randomized Plot" ?

Marcus Speh marcus at x4u.desy.de
Sun Jun 27 15:37:25 EST 1993


The Subject header says it all -- the background is this:
  after 20 years working in the hospital as an anaesthesiologist,
my father is returning to science. In his time, there were no
computers involved in medical data analysis, thus the question.
  I am a elementary particle theorist myself. We are going to publish
a study on the impact of certain drugs on patients -- the samples
taken are poor and don't give rise to "high statistics". [We have
convinced ourselves, though, that this is the case in any other paper
we've been looking at -- understandable since these aren't
'experiments' with animals in a cage] The function to be found lives
in a multi-parameter space, in fact, most of these extracted from
ECG and EEG data.
  Could a "randomized plot" be nothing else but a plot from 
random data, fed into a formula obtained from few data points? --
This would produce the desired high statistics, but the results
would be clearly to be taken cum grano salis, like any Monte Carlo
data. In physics, one doesn't "randomize" anything unless one is
forced to do so, ie. in absence of real data (with sufficient
statistics).
   I would prefer to know whether anyone reading this group -
preferably an actively publishing medical doctor - could provide
an answer. I am sure it is a trivial technical question.
   Maybe there also is more suitable newsgroup where to pose this
question?

	Thanks in advance,
	Marcus



More information about the Neur-sci mailing list