Ray Thomas writes:
>And, maybe a different question, is the concept of confidence level
>applicable where the main sources of error are likely to be
>non-sampling
>errors?
This seems to be saying that one should expect detectable differences in the
results of "errors" in the natural world due to random divergences from
law-like behaviour (canonical example: radio-active decay) and errors
arising from our imperfect measurements of the world.
To an amateur statistician like me this doesn't seem reasonable. Over to Ray
(and others).
Julian Wells
OU Business School
The Open University
Walton Hall
Milton Keynes
MK7 6AA
United Kingdom
+44 1908 654658
|