(no subject)

Hi,
Mark Taper, Subhash Lele, and I organized an ESA invited session around
this topic 
at the 98? ESA meeting in Baltimore MD. Which resulted in 
Mark L. Taper and Subhash R. Lele (eds): The Nature of Scientific
Evidence: Statistical, Philosophical, and Empirical Considerations 
University of Chicago Press. 

Ok shameless self promotion aside, I think there is a little more
to the argument that Edwards and Royall are putting forward, and how it
differs from the frequentist
approach. Once the likelihood (model) is specified, the sampling/design
is ancillary (Basu's Theorem),
than the comparison of two competing models is the likelihood ratio. The
p-value than adds the additional 
layer of a not very informative null model. The idea here is that the
sampling or design should dictate the model
and then competing models should be assessed by the strength of evidence
given the data. I think far more interesting
is how well does this translate to non-nested models, ie competing
mechanisms. 

I don't think that Royall and Edwards are alone, however the Bayesian
approach
is usually what is suggested as the alternative, and has its own can of
worms. 
There are some other interesting approaches as well, such as the notion
of model adequacy. 
I think in Ecology we are rarely interested in the false comfort a
p-value provides,
we are more often interested in the evidence for competing mechanisms,
or prediction
of future outcomes.

My 2c

Nicholas