|
Functional
dependence
When
Precision estimates for ore reserves was reviewed
by the Journal of Mathematical Geology, Dr George
S Koch Jr, coauthor of Statistical analysis of geological
data, recommended publication but Dr Roland Froidevaux,
a geostatistician with FSS International, would have none
of that. I questioned Froideveaux's findings and requested
an unbiased opinion. Professor Dr Robert Ehrlich, JMG's
Editor, entrusted that tough task.to Dr Margaret Armstrong,
JMG's Associate Editor, Editor of De Geostatisticis,
and coauthor of A study of kriging small blocks
(CIM Bulletin, October 1989). Armstrong's review, despite
its brevity, shows how geostatistical minds deal with
the concepts of.independence and dependence in classical
statistics. Here's but one tortuous thought of many.strung
together into an mind-boggling nine-line rant:
".
kriging variances are said to "violate the requirement
of independence" when they were developed precisely
to handle the case of (spatial) dependence ."
The
scientific facts are that kriged estimates violate the
requirement of functional independence in classical statistics,
and that kriging variances are a mathematical aberration.
In her sermon about Freedom of speech? (De Geostatisticis,
July 1992), Armstrong prevaricates about my first article
(The Northern Miner, May 1992) and ponders how
to deal with unpopular opinions:
"Does
the peer review process effectively deprive these people
of their freedom of speech by denying them the chance
to express opinions that run against the popular view?
Or is the peer review process just doing its job of
rejecting papers that do not back up their opinion with
scientific fact?"
|
