I checked and somebody went in and screwed up my fixes to the wikipedia page on Bayesian inference. I give up.

Posted by Andrew on 25 November 2011, 2:22 pm

## Recent Comments

- Daniel Lakeland on The curse of dimensionality and finite vs. asymptotic convergence results
- Daniel Lakeland on The curse of dimensionality and finite vs. asymptotic convergence results
- Nick Menzies on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Adam on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Anders Gorm on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Gilles on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Paul Alper on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Gilles on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Ben on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Bob on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Anoneuoid on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Jonathan (another one) on The graphs tell the story. Now I want to fit this story into a bigger framework so it all makes sense again.
- Jeff Walker on Low power and the replication crisis: What have we learned since 2004 (or 1984, or 1964)?
- Pren on Low power and the replication crisis: What have we learned since 2004 (or 1984, or 1964)?
- Aki Vehtari on The curse of dimensionality and finite vs. asymptotic convergence results
- Martha (Smith) on Let’s face it, I know nothing about spies.
- ojm on The curse of dimensionality and finite vs. asymptotic convergence results
- ojm on The curse of dimensionality and finite vs. asymptotic convergence results
- Chris Wilson on The curse of dimensionality and finite vs. asymptotic convergence results
- ojm on The curse of dimensionality and finite vs. asymptotic convergence results

## Categories

yeah, that’s the nature of wikipedia… do you know which revision is yours? since blog entries are on delay, i’m not sure which version to go back and read. (i’m not going to change it back. that’s not a battle worth fighting.)

Andrew’s revision 19th August: http://en.wikipedia.org/w/index.php?title=Bayesian_inference&oldid=445664320

Based on revision history, Andrew’s changes were removed by Gnathan87 15th September. Gnathan87 has been very active editor of Bayesian inference since July 2011.

jimmy:

Things will inevitably go forwards and backwards as different approached are tried out, particularly when an article needs major work. If something is reverted more than once, then it should be discussed on the talk page (as is now happening here). However, if nobody does this, it is assumed that everybody is happy. Please do not assume that people are necessarily being belligerent :)

I was in fact trying to take Andrew’s comments on board, in particular by trying to shift focus more clearly onto continuous rather than discrete examples. The reason I was bold enough to “revert” his edits was because the new description also explained how the “discrete” description applies equally to the continuous case.

In my view the old description is not understandable to readers new to the subject. Maybe the old explanation works for those who trust the process and can put the details to one side, but I have doubts about those coming to it for the first time. In particular, it does not explain how P(E) fits in to the probability space. What do P(E) and P(E|H) mean if the probability space is not defined?

One solution might be to additionally include a description of the method based on P(theta|E).

Lame. Edit it back and explain your reasons on the discussion page. Feel free to give up if you get stuck in an intractable dispute, but what makes Wikipedia great is that it is constantly being edited. Sometimes this is for the worse, but giving up after a single reversion is lame. You’re too used to people respecting your expertise (even if you’re generally right and know way more about bayesian inference than I will ever know).

I derive great value from Wikipedia for scholarship in my areas of specialization, but what kind of madman actually thinks “what makes Wikipedia great is that it is constantly being edited”? It is as if I said “what makes my personal paper based library great is that it is constantly being soiled by the acids and oils on my fingertips”. An individual edit may improve an article, but more edits does not mean more greatness.

Also, why must Prof. Gelman personally rescue his edits? In his two posts there have been several Wikipedia fans commenting how Prof. Gelman should spend his time, but his edits can be rescued by anonymous curators just as well as himself – even better than himself because the anonymous curators will probably have more free time and better knowledge of the social norms and conflict resolution/escalation policies of Wikipedia.

They completely removed/exiled mine on the strong likelihood principle. I too majorly give up: in philosophy of science entries, for example, it’s loaded with descriptions from radically analytic standpoints presented as giving general definitions presumably accepted by all.

Having done a bit of wikipedia editing in the past, I suspect that the best thing to do is to place some of your concerns about the article on the discussion page, signed with your name (maybe create a user page that identifies you as the Andrew Gelman).

I can’t say from experience that this works for an expert to swoop in and put their stamp on an article, but I can say that it helps to have a record of the motivation for the changes in order to help preserve them. Of course the other guy editing the page may be so stubborn that you can’t get anything done unless you’re willing to have a long protracted fight. =\

Just to follow up on this. Prof Gelman’s edits were restored (and there are subsequent edits on top of them): http://en.wikipedia.org/wiki/Bayesian_inference