To my comments:
> Based on this, most astronomers are not scientists. I guess it
> depends on what one means by "experimentation." Most statisticians
> distinguish between "observational studies" and "controlled
> experiments." The latter permits the use of Pasteur's word "prove."
> But, strictly speaking, the former does not.
Burgy responds:
> Possibly. I see astronomers experimenting the same way anthropologists
> and geologist experiment. While one cannot "run the universe twice,"
> all can posit a theory and then look for data to confirm/disconfirm.
> I'd call this experimentation.
As useful as this kind of activity is, what you describe here is a good
example of an observational study. The fact of prior positing a theory,
with confirming data, certainly adds strength to the activity. Still the
difference between observational studies and controlled experiments is
quite important when it comes to drawing conclusions. In this regard, you
might have a look at the first two (short) chapters of the excellent
elementary textbook entitled _Statistics_ by Freedman, Pisani and Purves:
http://www.wwnorton.com/college/math/stats/stats.htm.
This archive was generated by hypermail 2b29 : Sat Dec 29 2001 - 14:31:28 EST