[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]
From: Scott A Hartman [SMTP:firstname.lastname@example.org]
Sent: Friday, October 09, 1998 4:38 PM
To: Dinosaur Listserver
what the issue is here isn't about parsimony per se, but rather
application of parsimony to a data set as small as ours is in
paleontology (after all, look at how any times in the past the "most
parsimoniest" explanation has been overturned by new data).
The point about data size is important, but not necessarily a
defining issue, unless later data deviates significantly from existing data
and/or the interpretation of extant data stretches
It's actual statistical significance too much. There's no way to
Assure that a greatly limited database is not anomaly heavy
(represents an uncharacteristic population) without comparison to a more
robust database. The problem can come in with DEFINING a related database.
I crunch millions of bits of data every month,
On a variety of samples, but most are statistical samplings of
metrology. Out of self defense, I took some more advanced statistical
analysis courses. If you get bitten often enough,
You learn to examine your assumptions about the dog. :-) And
parsimony has nibbled on me a couple of times. So, William's ole razor can
sometimes seem to be a tad dull. One big cause of this can
Be hidden variables. Another special cause can be unexpected
interactions. Simplicity is not always so simple to arrive at.
The solution is (boringly) more data. More data & more data.