How do regional opinions on Twitter represent real-world attitudes toward climate change? A team of researchers will work to find out, thanks to a recent seed grant from the Social Science Research Instituteat Penn State.
In their study, the researchers will construct a Twitter-driven regional opinion heat map, using an opinion-mining pipeline and the social media platform’s geotags, and analyze whether the map represents regional attitudes toward climate change captured in national survey data. Then, they will examine how differences in location, time and occurrences of climate-related events shape people's opinions toward climate change. The researchers hope to gain a deeper insight regarding public opinion through the lens of social media.
"We chose climate change not only because it’s important, but also because the national survey data is available for research," said Ting-Hao (Kenneth) Huang, assistant professor in the College of Information Sciences and Technology and principal investigator on the project. "We have the survey results of public opinion regarding climate change in each area. We can use Twitter geotags to map social media opinions based on the data, giving us parallel assessment that we can compare with the survey results regarding public opinion about climate change."
Recognizing that Twitter users can be viewed as a biased sample of the entire population, Huang said that the team will employ statistical analysis techniques and large-scale survey data on climate change opinions to examine how representative social media opinions are.
"To really understand how social media opinions compare with opinions from survey results, we need to narrow it down to a level where we can quantitatively describe the differences," said Huang. "Once we do that, we can better understand the relationship between opinions in social media and those gathered by the survey."
The researchers also plan to use Twitter to study the dynamics of local attitudes toward climate change, and to determine the relationship between local climate-related events and opinions expressed in tweets.
"People can be influenced by climate-related events," said Huang. "We want to choose one event, like a certain hurricane, and look at the detailed data before or after the event. We will be able to observe whether or not people’s opinions would shift [after this climate-related incident]."
While the researchers will focus on climate-change opinions for this study, Huang said that the method could be applied to other widely debated hot-topic issues.
“We see modern media reporting certain hashtag trends of what people are discussing," he said. "That's a very important part of our digital life. To me, I want to see whether we can gain a deeper insight regarding public opinion through the lens of social media."
Huang is collaborating with Guangqing Chi, associate professor of rural sociology and demography and public health sciences in the College of Agricultural Sciences, and John Yen, professor of information sciences and technology, on the project. The seed grant funding was awarded by SSRI, in collaboration with the College of Information Sciences and Technology and the Institute for CyberScience. This is one of six University projects this spring to receive SSRI funding for developing innovative research programs using Twitter data. The team is pursuing external funding to further advance the project.