eMarketer reports that a recent study on traditional media’s use of blogs shows that 57.7% of respondents–US journalists –use blogs to measure sentiment.
Identifying Expressions of Emotion in Text (core study found here, registration required) is an intriguing “blog sentiment” study that identified targeted, emotion-laden words–“seed words”–and retrieved 173 blog posts from the Web that contained such. What the researchers found was that you can algorithmically determine the “emotive pulse” of a blog, or individual blog entries.
Once a product is developed around this method, one practical application of this is that you could quickly cull blog posts that target a certain emotion pertaining to a brand so as to determine positive/neutral/negative sentiment relative to that brand. This is particularly useful for brands concerned about maintaining a “real time” knowledge base as to their status in the blogosphere; thus, enabling brand managers to anticipate/preempt potential public relations crises.