What's the difference between Bayesian and classical statistics

Some interesting comments on the subject to be found here: What's the difference between Bayesian and classical statistics - Statistical Modeling, Causal Inference, and Social Science

I would say that Bayesians are the ones who are actually addressing the problems that most frequentists only think that they are...which I suppose is the same thing as Bill Jefferys' comment:

Since my background and training are in the physical sciences, I've noticed that all but the most sophisticated of my colleagues (that is, those that have learned enough statistics to be dangerous :0), think that a confidence interval is a credible interval. Which is natural, if mistaken.

Of course the IPCC proved that these entities are the same, back in the TAR...

Yet another comment on Schwartz

Via the perennial wikipedia editing wars, I find that yet another person has gutted that awful Schwartz paper:

On the diagnosis of climate sensitivity using observations of fluctuations by D. B. Kirk-Davidoff.

It appears to have been written a little later than the other comments, although overlapping such that Kirk-Davidoff didn't seem to know of our work before the initial submission (it is useful that ACP/ACPD has the old manuscript and reviews on line so the full history can be checked).

The comment is a more thorough investigation than we or Knutti et al could fit into the tight page limit of JGR comments, but I don't think there is anything materially different. As well as looking at the IPCC database (and concluding that the analysis method does not usefully indicate the equilibrium sensitivity of these models) Kirk-Davidoff investigated a slightly more complex model than the simple energy balance we used, and found that the method was useless for that too (or to be precise, in principle it would work given a long enough time series of observations, that is not available for the real climate system). I presume the reasons are much the same for the more complex model as for the simpler model we analysed, but there is probably no easy analytical solution in the more complex example.

There is no response from Schwarz in the review process (or as a separate paper, which I believe is possible through the ACP/ACPD system). I hope (and expect) that the editor, and indeed the author, contacted him about it. Of course he has already basically admitted it is wrong in his reply to us.

Odd, therefore, that the Schwartz stuff continues to get pimped on the Wikipedia page...

Plus ça change...

Good to know that the new "we will ban hereditary MPs" DPJ is holding firm on its beliefs.

Was that "four legs good, two legs...better?"

The creatures outside looked from DPJ to LDP, and from LDP to DPJ, and from DPJ to LDP again; but already it was impossible to say which was which.

[jules' pics] 9/23/2009 07:53:00 PM

mountain, originally uploaded by julesberry2001.

Clue 3:
Here's the view from the bookstore with the innernational turrism section.

But that was yesterday...

--
Posted By jules to jules' pics at 9/23/2009 07:53:00 PM

[jules' pics] 9/23/2009 07:22:00 AM

Our "Compact", originally uploaded by julesberry2001.

Clue 2:
The manual for the truck boasts than it can pull a boat, but then our Nissan Micra used to do that and it could also do 50mpg with a tandem on the roof. Happily we already have mud going up to the door handles. Today we head out into the wilds. I wonder if there will be more or less internet there.

--
Posted By jules to jules' pics at 9/23/2009 07:22:00 AM

[jules' pics] 9/22/2009 04:36:00 PM

In the bookshop, originally uploaded by julesberry2001.

Where are we?

Hint: Hired a small car, but they gave us a big truck that, apparently, does an "economical 28mpg".

--
Posted By jules to jules' pics at 9/22/2009 04:36:00 PM

The worst system, apart from all the others

Not democracy, but peer review, according to Sense About Science, via Nature.

The survey doesn't make clear how the authors were selected, which could make a difference if more prolific authors had a better chance of being picked. Or, indeed, if those who had suffered had actually left science and were unreachable. But still, there isn't much surprise there. Peer review works reasonably well on the whole, it can't always stop the misguided if they are determined enough but it can slow them down and on the whole I'm sure it improves the quality of the final papers, thus providing a benefit to the community. I have commented before that I like the idea of the EGU system where the review process is public but reviewers can remain anonymous. This is, however, only a tweak to the overall system.

IMO the more interesting debate in publishing relates more to the open access vs paywall question. For Climatic Change, I had the option of paying $3000 for my recent paper to be open access, so that's a benchmark for the value of the effort that authors, reviewers, and editors are donating to the profits of private businesses. (I know, the supporting editorial services are not completely free: the EGU manages it for about €25 per page.)

[jules' pics] 9/19/2009 12:03:00 AM

zuisenji stairs, originally uploaded by julesberry2001.

The weird thing is that this is not 2 photos stitched together. Sometimes temples invest their year-end extra-budget in a new staircase. Usually the old stairs are then cordoned off and revered, but not so at Zuisenji.

--
Posted By jules to jules' pics at 9/19/2009 12:03:00 AM

[jules' pics] 9/17/2009 09:00:00 PM

mug-cup, originally uploaded by julesberry2001.

Last day of school before the holidays! And I bought some photo-apps for my iphone to play with on the plane.

Meanwhile we try to capture the quintessence of the red higanbana, but fail because the amazing thing about them is the way the green stem shoots from bare ground and full bloom follows 2 days later. The leaves follow and feed the bulb over the winter. If you had a pet higanbana in your garden then a time lapse movie could be a lot of fun.

--
Posted By jules to jules' pics at 9/17/2009 09:00:00 PM

Friday is PARK(ing) Day

I'm all for a little bit of guerrilla gardening and reclaiming some small part of the area currently devoted to the demands of the automobile, so here's a quick plug for PARK(ing) Day UK - September 18th 2009.

Mind you, the suggested one day conversion of parking spaces seems a bit...weak. My own permanent PARK(ing) space can be seen here, in an old picture that dates back to before global warming really took hold (otherwise known as winter a few years back):

The viewpoint doesn't really show it off to best advantage - there are more plants off-screen to the right - but I can hardly do a better job now it's dark out. Anyway, being privately owned, I'm not sure if it counts, but it should. I do pay ¥25000 per month for it, but that is really a fiddle to get round JAMSTEC's silly rules over house rentals rather than a real parking space rental.

As an aside, I don't think there is actually that much need for guerrilla gardening in Japan. The climate being what it is, any unused space is reclaimed very rapidly (eg spot the greenness peeking though here). Also, people here are pretty keen on gardening and like to put pots out wherever there is a few inches of space - in fact that same residential road where the scrap metal dump is, is also lined with flowers, tomatoes, and chilli plants all through the summer. The problem is in really in preventing people from paving, repaving, and expanding over undeveloped areas at an ever-increasing rate.