In article <[EMAIL PROTECTED]>,
Christopher Tong <[EMAIL PROTECTED]> wrote:
>On 19 Sep 2000, Herman Rubin wrote:
>> >> This would exclude the application of probability to such
>> >> things as nuclear physics. While we have to use
>> >> observations to draw inferences, the probabilities of
>> >> interest are not those about the observations, but about
>> >> the underlying process.
>No, the probabilities of interest are both those about
>the observations and those associated with the underlying
>process.
I would disagree with this in most cases. The probabilities of
the observations are not likely to give insight, at least in
many cases. The process is what one makes assumptions about.
>One could take the view that predictions of probabilities of observable
>events are a means to an end: the continual development of
>better and better models of the underlying process. However, Hume's
>paradox guarantees that we will never be able to certify a given model as
>the "final one".
I agree that we will never even have a "correct" model. That
is one of the arguments against the use of p values.
>Alternatively, one could take the view that the continual improvement
>of models is a means to an end: improved calculations of probabilities
>of observed events. Science can and has delivered such calculations
>which have proven tremendously useful, in the form of applications, even
>to non-scientists.
This is engineering, not science. Science is not much driven
by it, and this is not adequately realized.
>A less partisan view is to respect the mutual feedback and interplay of
>improved models of underlying processes and improved calculations of
>observable events. In this view both types of probabilities are
>"of interest".
But the probabilities of interest in applications are not
likely to be those in the experiments. Those probabilities
are mainly tools, not items desired per se.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
[EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
