confidence levels.

[wolram]

http://arxiv.org/abs/astro-ph/0401198

How many cosmological parameters?
Andrew R. Liddle
Astronomy Centre, University of Sussex, Brighton BN1 9QH, United Kingdom
29 January 2004
ABSTRACT
Constraints on cosmological parameters depend on the set of parameters chosen to define the model which is compared with observational data. I use the Akaike and Bayesian information criteria to carry out cosmological model selection, in order to
determine the parameter set providing the preferred fit to the data. Applying the information criteria to the current cosmological data sets indicates, for example, that spatially-flat models are statistically preferred to closed models, and that possible
running of the spectral index has lower significance than inferred from its confidence.
---------------------------------------------------------------------
this is an interesting paper, I'm not sure if the A and B information
criteria are better than others, but this paper questions the
accuracy of CLs "confidence levels", does anyone have an opinion
on LIDDLEs paper?

[Nereid]

Nice paper! Very clear and readable.

I particularly liked the discussion of publication bias - we here at PF (A&C) have seen a more extreme version of this, in that a) many researchers like to 'blow their trumpet' and the journos they talk with like to write catchy articles, and b) PF members will pick and write posts about only the most arresting of these journos' articles.

Of course, Liddle is simply saying what I've been saying [;)] (though he says it with the backing of solid results from information theory, and writes a darn sight better than I can [*(] )

[Nereid]

Today's the 16th, and there have been 62 views of this thread, which is far more than the ~10 per post of a 'normal' thread.

What do all you other readers (PF members) think about the paper which wolfram so kindly posted here for us? Is it really cool, or is it not?? [?]

[wolram]

by NEREID.

I particularly liked the discussion of publication bias - we here at PF (A&C) have seen a more extreme version of this, in that a) many researchers like to 'blow their trumpet' and the journos they talk with like to write catchy articles, and b) PF members will pick and write posts about only the most arresting of these journos' articles.
---------------------------------------------------------------------
it is probibily worthy of a book on "how to avoid catchy science",
for some it is easy to spot media seeking researchers, but others
myelf included can fall into their influence, best advice wait
for reaction to publication, ask PF for opinion.

[selfAdjoint]

Nereid, it is truly cool. I particularly liked the part about publication bias, which certainly should be taken to heart! And I wonder about applying the same information criteria (AIC and BIC) to other areas with a lot of data. For example cognitive performance data and the g model.