On Thu, 29 May 2003 16:42:22 -0700, "David Heiser" <[EMAIL PROTECTED]>
wrote:
[ snip, much]

> A. W. F. Edwards  put together an interesting little book "Likelihood"
> (1972) that rather simply goes into the philosophic problems of "Maximum
> Likelihood" versus "Least Squares" 

I've owned a copy of Edwards for decades, I've tried to read 
it, and I don't think you have that right.
 
I've never heard that before, a suggestion that Edwards 
was *at all*  about likelihood versus least-squares. 
I certainly don't read it that way.  What I read is more like
what is suggested by the review/overview that I find at
Amazon.com --

"Dr Edwards' stimulating and provocative book advances 
the thesis that the appropriate axiomatic basis for inductive 
inference is not that of probability, with its addition axiom, 
but rather likelihood - the concept introduced by Fisher as 
a measure of relative support amongst different hypotheses."

In my personal, crude analogy,  I think of it as being a 
contrast between inference based on the cumulative
tail-probability [ probability]  versus inference based
on the ordinate of the curve [likelihood].


>                    In essence, "the twain shall not meet".
 - not a germane statement, as I can read it -

> Fisher was bothered by the "Maximum Likelihood" method, because the p value
> did not follow all the laws of probability. 
 - possibly -

>                                            Nevertheless, "Maximum
> Likelihood" is the method of research where data covariance structures are
> built and compared to theoretical models. 
 - Not relevant to the thesis of Edwards.  As I commented on 5/29,
folks have learned to frame some problems in ML terms, and 
they have developed relevant computer programming algorithms.
(Is it true that *all*  the Structural models programs are ML?)


>                                    Maximum likelihood also is the
> basis for data imputation methods. But all us simple statisticians just go
> ahead with least squares and are happy with the results.
> 

I thought a whole lot of imputation insisted on preserving
*means*  -- which, I thought, should imply least-squares.
Regardless, I don't remember people arguing about imputation
for reasons that were theoretical, rather than pragmatic.

-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to