The Tweeps Have Spoken: Twitter Comparable to Traditional Polling for Gauging Public Opinion
A new study proposes an innovative use of Twitter: replacing traditional polling as a means of of gauging public opinion.
Twitter, for its popularity, is still a bit mysterious. Despite increasing dramatically in use and social cachet, most folks still don't understand the real purpose of the social network. However, a research team at Carnegie Mellon may have discovered one of those uses: yielding public opinion statistics of comparable accuracy to public opinion polls, at vastly less expense.
According to a new report out of Carnegie Mellon University's computer science department, sentiments expressed via the millions of daily tweets strongly correlate with well-established public opinion polls, such as the Index of Consumer Sentiment (ICS) and Gallup polls. Data extrapolated from Twitter was strikingly similar to the data gathered from the traditional polls on topics like Obama's job performance, the job market, and the economy were strikingly similar - between 72% and 79% correlation. That may not sound particularly correlated - however, bear in mind that the ICS and Gallup polls themselves are only about 7% more correlated (86%) to each other.
"With seven million or more messages being tweeted each day, this data stream potentially allows us to take the temperature of the population very quickly," assistant professor and research team head Noah Smith explained. "The results are noisy [or distorted by superfluous information], as are the results of polls. Opinion pollsters have learned to compensate for these distortions, while we're still trying to identify and understand the noise in our data. Given that, I'm excited that we get any signal at all from social media that correlates with the polls."
For all this excitement, don't cancel your Gallup account yet - there were several areas where the researchers discovered that Twitter-derived data didn't correlate particularly well with the polls. Twitter mentions of Obama did track comparably to polling during the runup to the 2008 presidential election; however, mentions of McCain also tracked upward at the same pace, which did not correlate to polling data. This may be due to the demographic skew of the site - "Democrats as early adopters, eh?" blog Fast Company quips.
The data analysis methodology will require considerable tweaking before Twitter-derived readings of popular opinion would be as useful as poll-derived readings. "Improved natural language processing tools, as well as query-driven analysis and use of demographic and time stamp data... could increase the sophistication and reliability of microblog analysis," explains the study. However, the researchers are still very hopeful that Twitter posts could ultimately act as a "cheap, rapid means of gauging public opinion." Blog FastCompany asserts that this would be a means by which Twitter could monetize itself - a means that could perhaps outpace ad sales in monetization without changing the aesthetic of the site.
TrackBack URL: http://blogs.psychsterdata.com/cgi-bin/mt/mt-tb.cgi/94