Re: Media Testing for maclife.de
Re: Media Testing for maclife.de
- Subject: Re: Media Testing for maclife.de
- From: Richard Wagner <email@hidden>
- Date: Thu, 18 Sep 2008 14:12:56 -0700
On Wed, 17 Sep 2008 11:23:51 -0700 Chris Cox <email@hidden> wrote:
Again, the measurement itself is just a number (or set of numbers)
- by
itself, it cannot be wrong. If the meaning assigned to that number
does not
match expectations or a standard, then you have to do more
investigation.
It may mean a broken instrument (or a nearsighted assistant), or it
may mean
that you did not have a good experimental setup.
Or it means that the measurements are wrong.
You have a very convoluted way to look at the concept of a
measurement, and it's contrary to the way the term is used in real
life science. Measurements are frequently wrong. We commonly call
them "inaccurate measurements." They are the bane of the experimental
scientist.
It is an important point in sciences: just because the measurement
doesn¹t
match what you expected, that does not make the measurement wrong.
No, but it often does. Getting "correct" or "accurate" measurements
can be an art in itself. Take something as "simple" as measuring low
voltage signals with a USB board on a PC. For example, if you are
using a USB module with 16-bit resolution to measure a signal in the
+/-10 V range, the LSB value is 0.31 mV. If the module is
nonisolated and a transient voltage occurs in the electrical system,
your data could be off by *hundreds* of millivolts. Even in static
environments, your data could be off by tens of millivolts - huge
inaccuracies when you are measuring low-level signals. Ground loops
can also be a source of error. Depending on how you connect single-
ended inputs to the module, you can introduce ground-loop errors
that, when added to your signal and other ground potentials across up
to 5 meters of USB cable, can provide highly inaccurate
measurements. There is no "scaling factor" for these measurements -
the data are just plain wrong and the measurements are often
useless. It is necessary to work on the experimental setup until the
data are good and the measurements are accurate, and they give
expected, reproducible results. There are good data and bad data...
Usually it just means that there is something you failed to account
for in the theory that predicted the values.
No, usually it means there was an unaccounted for source of error in
the measurements.
If we threw away all results that didn¹t match expectations, much
of our scientific discovery would never happen.
If we tried to use your assumption that "the measurement itself is
just a number (or set of numbers) - by itself, it cannot be wrong"
then much of our scientific discovery would never happen.
I'm not sure what your training in bench science is, but your
arguments would have gotten me laughed out of my graduate program in
biophysics.
Anyway, this discussion is getting far away from the topic at hand.
I haven't seen any objective evidence that measurement errors account
for the results obtained by Uli.
--Rich Wagner _______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden