Power law distribution in news coverage…
I think Kevin Munden is correct in his conclusion regarding power law distribution in news coverage (from Jay Manifold’s A Voyage to Arcturus):
…the news media is every bit the interlinked, cross-connected, and self-referential network that the blogosphere is.
What makes one story bigger than another? I bet there are many in media who’d really like to know the answer to that one. I know scandal sells. Stories about kids sell. Celebrity and infamy Sell.
But I wonder why one story about an abducted kid takes off and has legs, while another pretty much identical one peters out after a day or two. Why the Niger Yellowcake story grows, while the bioweapon spewing drone aircraft story goes away.
What Munden did was create a distribution for links to stories from Google’s news page. The plot resembles a power law distribution, which argues that, across a wide spectrum of news organizations that Google aggregates, some stories have more “power” than others. In other words, what’s news across the network of journalism in all its forms is uniform.
At first this seems an intuitive situation, i.e. well, duh, news is news. But there’s something else here worth considering. If what is considered news creates a power law distribution, does that suggest that the influence and/or readership of certain media would also create a distribution similar to link distribution for weblogs (my comment here)?
I read an article recently (sorry, no link) that suggested The New York Times, as the so-called paper of record, exerts a vast influence on what’s considered news (including, especially, TV news). I happen to agree. Can such influence be plotted? Or consider the vast numbers of people who watch a given TV news program compared to the readers of a given newspaper. These numbers could be plotted and might create a power law distribution. But I think the plots would diverge, i.e. the influence (power) of The New York Times is out of proportion to is circulation.
That brings me back to the power law distribution of the blogosphere cited above. InstaPundit rules in influence and link numbers, but, in the grand scheme of all blogs and all readers, his blog draws only a fraction of the potential audience despite his influence.
I think I can begin to answer Munden’s first question: What makes one story bigger than another? Journalism is structurally biased toward certain ways of presenting information. That means it “sees” some things more easily than others and calls those things it sees news.
Further, I think Munden’s inquiry offers an argument for something I maintain is crucially important to understand about journalism: It is a network of (under-theorized) practice, e.g. the press applies a narrative structure to ambiguous events in order to create a coherent and causal sense of events. Common practice would account for the distribution of common news. So the influence of The New York Times comes from the perception by practitioners that its journalists practice best (despite recent bumps along its journalistic road).
Andrew,
Surely, some quantitative work has been done in the MassCom field on the agenda setting power of NYT?! I know that there has been some stuff in PoliSci done, especially on presidential rally effects during wartime, that uses NYT headlines as a major indicator.
James…yes, I’m sure there are. In addition to the headline indicators you mention, the NYT and WaPo are often used as the basis for content analysis in news stories. What I’m wondering is:
1- How could influence be quantified (we know it exists anecdotally)?
2- Might that quantification produce a power law distribution (I suspect it would)?
If this PLD story gets “legs” in the blogosphere, I might have to go look up this stuff
See also Clay Shirky’s The FCC, Weblogs, and Inequality, which asserts that a power-law distribution of media outlet audience sizes is inevitable without some pretty drastic government intervention.
What Kevin Munden stumbled across was the same kind of distribution among units of content (that is, news stories). Thanks to the First Amendment there is effectively no content regulation of US print media. Therefore the distribution of stories resembles that of the audience sizes of weblogs, which are also entirely unregulated and have essentially no startup costs. At least, that’s the lesson I draw from this.
The outstanding points of interest are:
1) How does this influence thing work? — probably a matter of networks of the sort touched on in _The Tipping Point_ and various other recent high-profile books about networking; but explicitly elucidating it, that is, finding that the news looks the way it does because people really do pay more attention to certain newspapers, etc, is still a worthwhile endeavor even if the theory is well in hand.
2) How do the structural biases of journalism, especially the narrative bias, feed the tendency for media to concentrate (in the absence of countervailing government action) into a few large, influential outlets and many smaller, uninfluential ones? Again, the theory may be well-understood, but a description of how a handful of newspapers and magazines have come to dominate political discourse in an immensely wealthy country of nearly 300 million people would be of great value. Am I enjoying the irony of ostensibly egalitarian journalists going about their business in such a way as to produce drastically unequal results? You bet I am.
I would also mention the influence of the associated press, which gets missed when people discuss centers of power in the industry. AP wields a huge influence over story coverage. But the NYT front page goes out the night before to many paper, and that surely has some influence.
Briefly noted…
The House of Representatives voted to roll back FCC ownership rules. Where is this heading? Who knows. Stay tuned. If I wanted to appear prescient, I would predict that there would be a compromise increase in ownership (say 40 percent,…
Jay… re: irony
Bryan… re: the AP. Good point. It’s an odd case in some ways because the AP, as wide network, does lead the news, but, as cooperative network, it also follows the lead of its constituent clients. So, I suppose that makes it an excellent news barometer. I’d be willing to bet we’d see the same distribution if we considered only stories run by the AP.
I, too, was going to comment on the wire service(s) and their influence on readership. How much NEW news are we receiving? How much is regurgitated (also considering that wire services not only feed TO newspapers, but feed FROM them).
As usual, I’m too late to post an original thought, but is an interesting point.