Communication and information science researches have developed a free online tool that offers a new way of looking at the news. "Reflext" provides a glimpse into how political events are being covered and discussed.
The Reflext site analyzes content from 10 major media outlets, more than 30 political blogs and recent presidential campaign speeches, using the linguistic concept of "selectional preference," which identifies associations among words. For example, words for liquids, such as "water," "coffee" or "oil" prefer specific relationships with verbs such as "pour," "flow," "freeze" and "splash." But what happens if we look at the terms associated with "Romney," "Obama," "Medicare"?
"Medicare" prefers terms such as "cut," "slash" and "end," but also appears with "reform," "expand" and "save," offering a hint about the perspectives and viewpoints surrounding issues of health care.
"How we think about political issues can shape the ways that we talk about those issues, often in very subtle and subconscious ways," said Eric Baumer, a postdoctoral researcher in the Interaction Design Lab headed by Geri Gay, the Kenneth J. Bissett Professor and Chair of Communication. "Examining how we talk about issues can provide insights into how we think about them."
The site allows users to select individual outlets such as Fox News, NPR or BBC News, and see what words are associated with a topic on each, or to look at all media together. From there, clicking on a word will show how it was used in context.
The site includes a how-to guide and information about the computer technology used.
"What we're looking at here is a fundamentally different way of reading," said Baumer. "Using tools such as Reflext allows us to see what's being said between the words and behind them. We can read not only individual articles but also patterns that permeate the broader media ecosystem."
The team that created the site includes students from the Departments of Communication, Information Science, Policy Analysis and Management, Linguistics, and Computer Science. The research is supported by the National Science Foundation's Social Computational Systems program and Cornell's Institute for the Social Sciences.