Unlike in the orchestra study they mention, the people doing the ratings have no way to differentiate between biased and neutral news articles, since they won't

Unlike in the orchestra study they mention, the people doing the ratings have no way to differentiate between biased and neutral news articles, since they won't be familiar with the underlying facts involved. The only thing that can really be judged by reading an article is how eloquently it was written. It seems pretty reasonable to use brand perception (in lieu of anything else) to judge bias, particularly in political articles.