Anecdotal facts that are not scientifically sampled.
Uh, with a sample size of 400,000, you get a statistically valid result without "scientific sampling", unless he has "cherry-picked" a subset of his entire data set to present--which I doubt is the case.
Anecdotal facts that are not scientifically sampled.
VA's web stats are scientific enough to be pertinent to the topic of the article and the subsequent discussion.
If you do some research on web development sites, you'll see dozens of other data collections with similar results, (i.e. MSIE is being used by well over 90% of internet users.)
The stats are important to developers because it is a PITA to write extra code for multiple browsers using older standards. Many web developers are ignoring the needs of browsers older than MSIE 5.x, and off-brand browsers that go tilt on XHTML, and they can do that with little risk to their target market.
Just because someone tracks and uses the data doesn't mean they are a fan of Microsoft; it just means they are paying attention to their market, and using that information to stay in business.