The State of the Union address is one of the few times each year when a large percentage of Americans reliably pay attention to politics. Once upon a time, as legend has it, things were different: most Americans tuned into Walter Cronkite in the evening or picked up the morning newspaper, which covered matters of national and international importance, like politics, foreign affairs, and business developments.
If analysts at Microsoft Research are correct, a startling number of American Web users are no longer paying attention to the news as it is traditionally defined. In a recent study of “filter bubbles,” Sharad Goel, Seth Flaxman, and Justin Rao asked how many Web users actually read the news online. Out of a sample of 1.2 million American users, just over fifty thousand, or four per cent, were “active news customers” of “front section” news. The other ninety-six per cent found other things to read.
The authors defined an active news customer as someone who read at least ten substantive news articles and two opinion pieces in a three-month period—if you remove the requirement of reading opinion pieces, the number of news readers climbs to fourteen per cent. The authors studied U.S.-based Web users who, between March and May of 2013, accumulated a total of 2.3 billion page views.
News can be a vague category; the authors defined by collecting news sites with appreciable traffic (the New York Times, the Huffington Post, and Fox News), blogs (Daily Kos and Breitbart), and regional dailies (the Seattle Times and the Denver Post). Using “machine learning” algorithms, the authors separated what, based on word usage, they considered front-section news from the other content on news sites, like sports, weather, life style, and entertainment. What’s left is the narrow, classical news article, about, say, the State of the Union, as opposed to one about the latest adventures of Justin Bieber or Farrah Abraham.
Various influences shaped the study. The data was collected only from Internet Explorer users (who, the authors say, tend to be slightly older), and it represents only those who agreed to make their Web-browsing history available. Additionally, just because people don’t surf news Web sites doesn’t mean that they don’t get news from other sources, like physical newspapers, talk radio, Twitter, “The Colbert Report,” or the evening news.
That said, the sample size, 1.2 million, is impressive—far greater than that of a typical survey. And the number of people whom the study shows to be paying attention to the news online is consistent with the low ratings of cable news during the same period. Also, as opposed to relying on what people said they did, the Microsoft researchers drew on a record of what they actually did, which is significantly different. In a 2012 Pew survey, for example, thirty-nine per cent of people said that they had read news online the day before. The difference between the two numbers—fourteen and thirty-nine—may, in part, reflect different definitions of “news.” (The Pew survey did not define the term.) And, of course, what people like to think they do is often different from what they do.
Assuming that Microsoft’s numbers show a real phenomenon, though, they do introduce some perspective. Journalists and political junkies often presume that everyone cares about politics all of the time. But the fourteen-per-cent number makes it seem more like a hobby or a subculture, something like the N.H.L. or Nascar—a deep obsession for some of members of the population but of limited interest to anyone else, unless something extreme happens.
To be sure, twelve to forty-two million potential readers is a respectable audience; it’s more than that of mixed martial arts, say, even if it’s not at the level of N.F.L. football. But, mainly, it suggests that attention to politics, once a basic mandate of citizenship, is now an entertainment option, in fierce competition with other forms of entertainment. Politicians’ awareness that they don’t have a guaranteed audience may also account for the increased use of reality-TV strategies in politics. We can see the congressional shutdown, in part, as an effort to move the numbers.
The figure is also important to the concept of filter bubbles, the main subject of the Microsoft paper. Commentators and journalists (and, of course, Barack Obama) have long bemoaned the division of America into highly polarized ideological camps, said to bereinforced by online filtering. But it seems that the most important filter bubble is the one that could be labelled, simply, “ignore everything.” It’s the bubble filled with people who, so long as the country remains basically stable, pay no attention to partisan politics.
The number may also help us understand why a relatively small number of motivated people can have such a significant effect on American politics and policy. For better or worse, the number of people in this game is pretty small. Bottom line: if you can get one per cent of the population vaguely interested in something nowadays, that’s huge.
By Tim Wu, The New Yorker