Tom Webster, writing and speaking

Can the U.S. Mood Be "Inferred Through Twitter?"

Added on by Tom Webster.

This report claims it can, and the New York Times ran with it. Once that happened, the Twittersphere started tweeting it (presumably, making Twitter users "happy.") Anything to this? Well, I'd point out that as of earlier this year, only about 7% of Americans were even on Twitter. The data was also collected from 300 million Twitter messages stretching back to 2006. The first time we ever added Twitter usage to our tracking surveys was in late 2007 - when Twitter usage was essentially a rounding error and not even representative of .1% of Americans. So this data, while colorful and interesting as a snapshot of Twitter, can hardly be billed as representative of the mood of the nation. Still, I'm not going to beat up what is a fascinating look at the mood of persons who have updated their status on Twitter (which is not the same thing as Twitter users.) Where such a study becomes interesting to a researcher like myself would be if a similar chart could be constructed of the actual "National Mood," and then compared to this data. If you had both datasets, then you would not only have the true National Mood (you don't need the Twitter data for that), you would also have a valuable means of calibrating the tendencies of Twitter users by examining where they differ, and where they resemble, the 90+% of Americans who don't use Twitter. For instance, our last data on Twitter users revealed that they were more optimistic about a U.S. economic recovery - but is that due to their own economic circumstances, or are Twitter users actually generally more optimistic people? For social media researchers, that is pretty valuable stuff.

So, "pulse of the nation," as the article claims? No, but pulse of Twitter, yes. Comparing the two would be a seriously interesting study, but one I'm in no mood to conduct.