Wikipedia approaches its limits - guardian.co.uk Options

[Newsfeed]

Guardian: Wikipedia approaches its limits

[John Limey]

Not sure if it's possible to accurately measure POV. One solution is to simple publish a list of problems inherent to articles, listing biases etc. It would be something like Rational wiki's articles such as one the on Citizendium, but perhaps a lot more detailed then what they skim over:

http://rationalwiki.com/wiki/Citizendium#Crank_magnet

If someone could figure out how to measure the POV-pushing and ownership of various articles, that would go a long way toward highlighting the inherent unreliability of Wikipedia over a multitude of subjects. But measuring POV is problematic. The first step, I venture would be to simple compile a list of biased and/or battleground articles, from Armenia (T-H-L-K-D) to Ireland (T-H-L-K-D) and from Evolution (T-H-L-K-D) to PETA (T-H-L-K-D). Some statistics on those articles and their owners might be illuminating.

It's completely impossible to measure POV in any reasonably objective way. While it is easy enough to find qualified experts to check the accuracy of an article, it's nearly impossible to find anyone who would be universally considered a qualified expert to check the neutrality of an article. With a few notable exceptions, the articles that are the subject of the most POV-pushing on Wikipedia are the subject of the most POV-pushing in the real world. Beyond that, measuring bias is an inherently subjective task.

On the other hand, "ownership of various articles" could be measured with a method similar to the PARC study. Just take a sample of articles, and see how likely an edit made by someone who is not a major contributor to the article is to be reverted and you already have a fair measure for ownership.

[Kelly Martin]

On the other hand, "ownership of various articles" could be measured with a method similar to the PARC study. Just take a sample of articles, and see how likely an edit made by someone who is not a major contributor to the article is to be reverted and you already have a fair measure for ownership.

Now this is an interesting idea. However, I'm not sure how to define "major contributor".

It might be interesting to have a metric that related the number of editors making nonreverted contributions to an article to the number of editors reverting contributions to the same article. It seems to me that an "owned" article will have a relatively small number of editors making nonreverted contributions. Contrariwise, an "unowned" article would have a larger number of editors making nonreverted contributions. The overlap between the set of nonreverted contributors and reverters would also be interesting.

I really should download a history dump sometime and play with it; problem is that I don't really have enough disk space, or processing power for that matter.