It’s been said by a number of commentators that blogging depends for its existence on the mainstream media. In terms of news reporting, blogging contributes little original content, say the critics, despite the odd scoop driven by ideological agendas and personal vanity.

So far that has been opinion backed up by often sketchy and selective evidence. Now we have firmer proof in the form of an algorithm which tracks news topics, ideas and memes across the Internet, and analyses persistent temporal patterns in the news cycle. The model developed by Jure Leskovec, Jon Kleinberg and others at Cornell and Stanford Universities reveals a typical time lag of 2.5 hours between peaks of attention to a phrase published in news media and blogs, respectively. It also shows that only 3.5% of the stories studied originate in the blogosphere.

In the case of media-to-blog information transfer, the researchers see divergent behaviour around an overall peak, and what they describe as a “heartbeat” pattern in the hand-off between the news source and blog reaction. The algorithm was tested on some 90 million journalistic news reports and blog posts published in the three months leading up to the US presidential election last year, and the results will be presented next month at the Knowledge Discovery and Data Mining conference in Paris.

Journalism still rules, even if not OK.