On Wed, 22 Jun 2005 15:31:07 -0400, "Walter Maxwell"
wrote:

I compared readings of
adjacent frequencies for two different heights where the retrograde occurs and
found differences in original R values that I can explain only in the
possibility of different degrees of soil wetness, because the measurements were
not all taken on the same day.

Hi All,

This is a life's lesson in the value of context and measurement, as
well as in the discipline of taking notes. Walt's memory suggests a
reason for the perturbation observed in the data, and it is not
unreasonable. I would suggest that there is some (however slight)
likelihood that the correlations may be backwards in that most of the
days followed rain, and these perturbations were on dry days.

Before or after is not the issue. Before or after is a matter of
separability which is more important. With analysis, Walt's
conjecture can be tested against the data and what it reveals about
the impact the ground's proximity had on the antenna.

His data, either way, already supports that ground is measurable
within the data that falls outside of the spread of noise and error.
Even if Walt slipped an instrument reading or injected statistical
noise, he did it so consistently that he was always in error in the
same direction (this is called systematic error). This may impact the
accuracy of the final answer, but it does not impact the thesis'
general conclusion. What is more, even if such mischance occurred
(and I doubt it), it is recoverable with one or several cardinal
measurements to correct the earlier bias.

This round of discussion also reveals that bad data is as good as good
data. Those who discard results and tailor their reports stand a good
chance of not discovering how to fix their problems when they are
shown to be in serious error (which is to say they probably rejected
good results). I pointed this threat out in another thread that
linked to exhaustive ground data that showed hills composed of fresh
water.

73's
Richard Clark, KB7QHC