Hot Story to Has-Been: Tracking News via Cyberspace By PATRICIA COHEN

August 5, 2009
Hot Story to Has-Been: Tracking News via Cyberspace By PATRICIA COHEN

Like a lot of new ideas, Media Cloud started with a long-running argument among friends. Ethan Zuckerman and a handful of his colleagues at the Berkman Center for Internet and Society at Harvard Law School found themselves in endless disputes about the mainstream media and newer digital variations. Who sets the agenda? How is public debate shaped? What topics are covered or ignored?

Anecdotes favoring one side or another were as plentiful as pop-ups, but a comprehensive and reliable database that could track the daily rhythm of the news cycle over time and was available for public use didn’t exist. So Mr. Zuckerman and others at Berkman decided to create one.

The result is Media Cloud, a system that tracks hundreds of newspapers and thousands of Web sites and blogs, and archives the information in a searchable form. The database, at mediacloud.org, will eventually enable researchers to search for key people, places and events — from Michael Jackson to the Iranian elections — and find out precisely when, where and how frequently they are covered, said Mr. Zuckerman, whose official title is senior researcher, though he acknowledges that a more accurate label would be computer geek and international development specialist. (At the moment only a small sample of Media Cloud’s tools are on the public Web site.)

The findings, which can be graphed or mapped, can demonstrate the evolution of a report and variations in coverage. Users get to “do the fun part, which is analyzing the data,” Mr. Zuckerman said, “while we do the hard part of this, which is collecting it.” Eventually users will be able to compare the top 10 news events covered by Fox News, The Atlanta Journal-Constitution and the BBC, for example, or chart the terms that appear most frequently in The New York Times, compared with leading blogs, or create a world map showing which countries receive the most media attention, or follow the path of a particular report to see if it dominates the news or dies out.

For the past decade or so, many researchers have used link analysis to figure out how information spreads, said Yochai Benkler, a Harvard Law School professor at Berkman who has been involved in creating Media Cloud. You could identify which Web sites were linked to most frequently and infer whose sites were most influential. But researchers have pretty much squeezed all that they can from that approach, Mr. Benkler said. Although Media Cloud is still in its early stages, it is among “the next generation of tools that actually look at what people are saying,” he said, adding, it is “a better microscope.”

There are other kinds of media trackers. Cornell University researchers, for example, have developed MemeTracker, which maps the daily news cycle by grabbing repeated quotations from one million online sources. (A meme is anything — an idea, a phrase — that spreads by imitation from one person to another.)

Its graphs, which can be viewed at memetracker.org, display the reports that are competing against one another for attention on a given day, as well as those that have staying power or quickly disappear. A recent paper on MemeTracker’s experience during the presidential campaign was hailed by experts as a landmark piece of work.

Pew Research Center’s Project for Excellence in Journalism offers a news-coverage index, which is laboriously compiled by having 14 people sample leading reports produced by 55 outlets. Media Cloud is much less exact, Mr. Zuckerman said, but it can automatically scan hundreds, and eventually thousands, of sources.

Amy S. Mitchell, deputy director of Pew’s journalism project, said Media Cloud “offers the public a great opportunity to play around with looking at a wide swath of media at more of a surface level.” But, she added, it cannot really capture the nuances of the news agenda of the news media. “There are certain things that computer algorithms cannot do that individuals can,” she said.

Since every method has virtues and drawbacks, she added, “I think there is tremendous value in having both approaches.”

What Media Cloud offers that no one else does is a tool anyone can use to answer all sorts of questions about the media landscape. One topic that Mr. Benkler and Mr. Zuckerman have long been debating is whether the Internet has helped open up the public sphere to more voices, or whether it just serves as an echo chamber, simply repeating information and views that the mainstream media already circulate.

Mr. Benkler is using Media Cloud to test his theory that digital media is widening the circle of voices somewhat. Sites that he characterizes as “one link out” from the most visible (like The Huffington Post, Talking Points Memo and Instapundit) are entering into the conversation, he argues.

Who has the power to place an idea on the national agenda is another question that Mr. Benkler said Media Cloud could help answer. For instance, how is the conversation about the recession and the financial crash shaped? Using some of the database’s more specialized tools, Mr. Benkler investigated who first floated the idea for a temporary takeover of the financial system by the government, as was done in Sweden in the 1990s.

Paul Krugman, a columnist for The New York Times, first raised the idea in September 2008, but it was a cluster of influential economic and political bloggers like Brad DeLong and Matthew Yglesias who kept the idea alive. After the subject disappeared for a couple of months, the bloggers then resurrected it early this year as Washington began discussing the details of a bank rescue plan. To Mr. Benkler, this preliminary evidence suggests that the network of public media has given a voice to some people who in the past may have had useful ideas but were, as a practical matter, unable to inject them into the national conversation.

“If you’re actually trying to map where an idea starts and how it moves through the public sphere, you need a database like we’re developing, with time-stamped data,” Mr. Benkler said, explaining that services like Google and Lexis/Nexis are not as comprehensive or do not provide that level of detailed information. Media Cloud also enables more fine-grained analyses by examining language and context. “How does rhetoric change over time, and what’s the role of the Internet and the mainstream media in that?” Mr. Zuckerman asked.

Some of his colleagues, for example, have been tracking the frequency of the words bailout and stimulus to pinpoint when one term overtook the other. Media Cloud mapped the results to show how the term bailout, used constantly in the news in the fall, eventually gave way to the word stimulus after President Obama took office. The results were graphed, illustrating precisely when the two lines crossed — where, as Mr. Zuckerman would say, “one meme took over from the other.”

Media Cloud’s founders have put out an open call on their Web site for research ideas. The system provides a platform, Mr. Zuckerman said, on which others can build using their own kinds of tracking software. The point is to start with facts rather than impressions.

As Mr. Zuckerman noted, a lot of anecdotes don’t necessarily add up to the truth.