Today, I stumbled upon a NYT review/summary post by Steve Lohr of a recently published academic paper titled ‘Meme-tracking and the Dynamics of the News Cycle‘ (by Jure Leskovec, Lars Backstrom and Jon Kleinberg of Cornell University). The paper analyzed how quotes/text-snippets from Obama, McCain and other important figures surged and spread across 1.6 million mainstream media sites and blogs during the last three months of the U.S. presidental race in 2008. They found, that traditional news outlets led and the blogs followed with a time lag of about 2.5 hours. Only about 3.5 percent of story lines originated in the blogs and later made their way to traditional media.

The full paper can be read or downloaded as a pdf-file here:

http://memetracker.org/quotes-kdd09.pdf

Additionally, the research team set-up an interactive display of their findings on http://memetracker.org.

Our thoughts on this paper:

We agree with Eric Horvitz, a researcher at Microsoft and president of the Association for the Advancement of Artificial Intelligence, in as far as this work is a ‘landmark piece of work on the flow of news through the world’. However, we are somewhat more critical about the findings of this seminal work. The paper does not discuss potential biases from selecting an event that is intensely covered by the big media players. Such biases may result from several particularities of such an event, maybe the important one being that big media allocates a lot of resources to cover and analyse such big events, a lot more than on regular and surprise stories. As there is just about every little aspects covered by big media, there is little room for bloggers to trigger yet ‘uncovered’ issues and stories. Finally, as big media companies have their correspondents at the campaign events or broadcast the most of the important events during a campaign, they are the primary news source for others covering the event, which gives them a little head start. Nevertheless, the paper provides a break-through at analysing how news spreads across news outlets on a large-scale, quantitative basis.

Update: Here is another critical review of the paper worth reading:

http://www.niemanlab.org/2009/07/in-the-news-cycle-memes-spread-more-like-a-heartbeat-than-a-virus/