Staticticians are exposed to a wide range of measurements, so perhaps someone
can help with my off-topic question. I direct it to allstat-l because I am at a
loss for who else to ask!
My problem is that I need to measure change in a property that is hard to
measure. I would like to measure another property, which is correlated with the
first, to estimate change in the first, but I also wish to avoid measurements of
the total of the first property to derive my regressions. I am thinking that I
should be able to estimate amount of change by setting up a calibration
experiment in which I manipulate the first property to determine values of the
second property within the range of change, and thereby avoid estimating the
totals entirely.
Can anyone provide examples from any field where something like this is done?
Or perhaps a source where this query might be better posted?
=======================================
Stephen G. Bousquin
[log in to unmask]
=======================================
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|