The notion was that a pattern of articles that tend to support the same group should raise our estimate of a bias in favor of that group.
All of them could be perfectly accurate, incidentally - consider a journalist who investigates 100 stories for every one he publishes and only publishes those that are favorable; while none of the articles would contain actual falsehoods, clearly this is a bias.
Whether that's the case here is a much deeper question that I can't answer from my extremely limited skimming of the available media and this thread.
All of them could be perfectly accurate, incidentally - consider a journalist who investigates 100 stories for every one he publishes and only publishes those that are favorable; while none of the articles would contain actual falsehoods, clearly this is a bias.
Whether that's the case here is a much deeper question that I can't answer from my extremely limited skimming of the available media and this thread.