Some researchers in Austria are trying to tackle a very tough issue, blog credibility ranking. Unfortunately (if this blog article is credible) they are building their house on a very shaky foundation:
The proliferation of widespread Internet access has enabled everyone and their dog to start a website, but not every one is filled with what some of us would describe as “credible” information. That’s why some researchers are attempting to create software that can analyze Web content and automatically rank it to help out those who can’t quite decide for themselves.
Researchers at the Austria-based Know-Center are working on a program that analyzes the language used on blogs in order to rank them as highly credible, having average credibility, or “little credible.” The code looks at the distribution of words over time, and compares blog topics against articles from mainstream news, which are apparently weighted as being more credible.
There are a whole host of reasons why using the mainstream media as a credibility benchmark is a bad idea, but the biggest reason is that the media really does a poor job in general of getting facts right. That’s not a criticism and really no one should expect any different.
The mainstream media generally gets things wrong simply because the content is generated by people that are generally not subject matter experts on the things they are writing about. They are experts at writing and (usually) journalism. Yet they are called upon to write about a vast universe of subjects that they are barely familiar with.
If you don’t believe me think back to the last time you read an article in the media that concerned a subject you were an expert at. Did you feel that the article accurately portrayed the crucial issues? Did it do it better than the dozens of blog entries you might have read on the same subject?