On May 9, 2009, at 12:35 AM, Sabri Oncu wrote:
Shane:
That astrology's theoretical structure finds no place among the
theoretical
structures of the "respectable" sciences in no way negates the fact
that it is, in principle, scientific. Empirical testing could
invalidate--or (perish the thought) confirm it.
But then, there is the data issue: should it be time-series,
cross-sectional or panel (that is, time-series/cross-sectional)?
If it is time-series or panel, what should be the observation
frequency: yearly, monthly, hourly, every other secondly or shorter or
way longer than yearly?
If it is cross-sectional or panel, how large should be the
cross-section: five persons, 5,000 persons, 5,000,000 persons, the
entire human population currently alive and lived in the past?
How about fitting a two-dimensional Fourier Series to an infinitely
long-infinitely broad data?
The fit would be perfect, but would it explain or predict anything?
Start with something much simpler. Take fifty matched pairs--one a
lottery winner, one a victim of a fatal automobile accident. Give
birth day/time/place and event day/time place for each to 200
professional astrologers (sample size=10,000). Tell them to identify
which was which in each case. Test the percentage of correct
identifications for statistical significance.
Shane Mage
This cosmos did none of gods or men make, but it
always was and is and shall be: an everlasting fire,
kindling in measures and going out in measures."
Herakleitos of Ephesos
_______________________________________________
pen-l mailing list
[email protected]
https://lists.csuchico.edu/mailman/listinfo/pen-l
