Twitter analysis shows climate skeptics shout loudest

Back in September, the first of three sections of the massive Intergovernmental Panel on Climate Change (IPCC) report on climate change was released. Unsurprisingly, it said some ugly things: that “warming of the climate system is unequivocal”; that the seas are rising at a higher average rate than that “during the previous two millennia”; and of course, that “human influence on the climate system is clear.” The response to such a touchstone event in climate science and policy could be seen in a number of ways, from media coverage to politicians’ statements, but one research group has now looked to the social masses—Twitter, specifically—for another measure of reaction.

“Some scholars argue that the climate change debate has become polarized between those classified as convinced of anthropogenic reasons for climate change and those skeptical of those reasons,” wrote study authors led by Warren Pearce, of the Institute for Science and Society at the University of Nottingham in the U.K. “Was there a similar polarization when people tweeted about the IPCC report?” They looked for tweets featuring the term “IPCC” around the release of the report in September and October 2013, and examined the connections established between those people, and if clear polarized sides could be identified and described. Results appeared in the journal PLoS One.

Among more than 61,000 tweets considered to contain “original content”, they whittled down the user population to those with the most active connections, yielding a sample of 239 users. Among those, the investigators found that about half of the highlighted users (49 percent) were “supportive” of climate science, scientists, the IPCC, or action on climate change. Meanwhile, only 26 percent were unsupportive, with another 22 percent simply tweeting out links without comment and classified as “neutral.” (In case you did the math: the final 3 percent had “conversational connections” with other main users but did not tweet specifically in this data set’s time frame.)

The interactions between these groups yield some interesting insights. First, they found that the “unsupportive” group initiated far more conservations than the other groups; in other words, they were the most insistent tweeters out there when it came to the IPCC. Here’s a visualization showing just that, with the purple nodes representing “unsupportives”, red representing “supportives” and green representing “neutral”:

Connections between twitter users on the IPCC report

As the big purple circles tell you, those skeptical of the IPCC’s findings, or perhaps of the need to do something about climate change, were talking the loudest. But who were they talking to? Using a statistical test, the researchers first proved the obvious point that we all tend to have more conversations with like-minded people. But they also showed that skeptics tended to have more conversations with themselves, compared with those supportive of the IPCC. The interactions with those neutral folks also were stronger with supportive tweeters than unsupportive ones. Importantly, in spite of those differences, there still did exist connections between all types of tweeters, showing the echo chambers aren’t 100 percent soundproof.

There is other information buried in our tweets as well, including some interesting insights based on hashtags. A “community” of users predominantly based in Australia, for example, involved more politicians than other geographical communities, perhaps related to that country’s elections and controversy surrounding a carbon tax around the time of the IPCC report’s release. The complicated interactions between all these users, though, led the researchers to conclude with some “cautious optimism” for improved communication across ideological lines. – Dave Levitan | April 22 2014

Source: Pearce W, Holmberg K, Hellsten I, Nerlich B. (2014). Climate Change on Twitter: Topics, Communities and Conversations about the 2013 IPCC Working Group 1 Report, PLoS One, 9 (4) e94785. DOI: 10.1371/journal.pone.0094785

Image:, Iculig