Sunday's New York Times book review included an essay by federal circuit court judge and law professor Richard Posner purporting to explain why “the conventional news media are embattled.” But the article is full of unsupported claims.

Posner made the common conservative argument that since there are more liberals than conservatives among journalists, the news must have a liberal bias. “Fourteen percent of Americans describe themselves as liberals, and 26 percent as conservatives,” Posner wrote. “The corresponding figures for journalists are 56 percent and 18 percent. This means that of all journalists who consider themselves either liberal or conservative, 76 percent consider themselves liberal, compared with only 35 percent of the public that has a stated political position.”

Posner didn't cite the source of these figures, but they contradict several reputable surveys of journalists. In fact, while the number of journalists who call themselves conservative is smaller than the proportion in the general population, they are not heavily outweighed by liberal journalists, because a clear majority of journalists describe themselves as moderates. As a 2004 survey by the Pew Research Center for the People and the Press found:

About a third of national journalists (34%) and somewhat fewer local journalists (23%) describe themselves as liberals; that compares with 19% of the public in a May survey conducted by the Pew Research Center. Moreover, there is a relatively small number of conservatives at national and local news organizations. Just 7% of national news people and 12% of local journalists describe themselves as conservatives, compared with a third of all Americans.

The remainder -- 54 percent of national journalists and 61 percent of local journalists -- called themselves moderates. But Pew itself cautioned against making too much of the relatively high numbers of liberals among journalists. The report's authors wrote, “But what does liberal mean to journalists? We would be reluctant to infer too much here. The survey includes only four questions probing journalists' political attitudes, yet the answers to these questions suggest journalists have in mind something other than a classic big government liberalism and something more along the lines of libertarianism. More journalists said they think it is more important for people to be free to pursue their goals without government interference than it is for government to ensure that no one is in need.”

Another highly regarded study, the 2002 American Journalist Survey, found that 37 percent of journalists called themselves Democrats (just a few points above the proportion of Democrats in the general population), 18.6 percent said they were Republicans, and 33.5 percent were independents. Thus while the journalistic community is not an ideological or partisan mirror of the public, it is more due to the absence of conservatives and Republicans than the dominance of liberals or Democrats.

Further, Posner's use of the formulation “of all journalists who consider themselves” is a convenient way to produce a high percentage of purported liberals-- 76 percent! -- but excluding moderates obscures more than it illuminates. One would get nearly the same result if the actual figures were that 3 percent of journalists considered themselves liberal; that 1 percent considered themselves conservative; and that 96 percent considered themselves moderate.

But at least in this case, Posner attempted to offer something resembling evidence for his assertions, which cannot be said for most of the essay. “The mainstream media are predominantly liberal -- in fact, more liberal than they used to be,” Posner wrote. “But not because the politics of journalists have changed. Rather, because the rise of new media, itself mainly an economic rather than a political phenomenon, has caused polarization, pushing the already liberal media farther left.” Posner offered no evidence for his assertion that the “already liberal media” has moved “farther left.” In order to explain this nonexistent phenomenon, he offered an overly simplified view of the consequences of barriers to entry on the ideological makeup of news organizations:

To see what difference the elimination of a communications bottleneck can make, consider a town that before the advent of television or even radio had just two newspapers because economies of scale made it impossible for a newspaper with a small circulation to break even. Each of the two, to increase its advertising revenues, would try to maximize circulation by pitching its news to the median reader, for that reader would not be attracted to a newspaper that flaunted extreme political views. There would be the same tendency to political convergence that is characteristic of two-party political systems, and for the same reason -- attracting the least committed is the key to obtaining a majority.

One of the two newspapers would probably be liberal and have a loyal readership of liberal readers, and the other conservative and have a loyal conservative readership. That would leave a middle range. To snag readers in that range, the liberal newspaper could not afford to be too liberal or the conservative one too conservative. The former would strive to be just liberal enough to hold its liberal readers, and the latter just conservative enough to hold its conservative readers. If either moved too close to its political extreme, it would lose readers in the middle without gaining readers from the extreme, since it had them already.

But suppose cost conditions change, enabling a newspaper to break even with many fewer readers than before. Now the liberal newspaper has to worry that any temporizing of its message in an effort to attract moderates may cause it to lose its most liberal readers to a new, more liberal newspaper; for with small-scale entry into the market now economical, the incumbents no longer have a secure base. So the liberal newspaper will tend to become even more liberal and, by the same process, the conservative newspaper more conservative.

Although reasonable in theory, Posner's analysis was unrealistic. First, he offered no evidence that the cost of running a newspaper has changed meaningfully in recent years. Further, his example of the two-newspaper town notwithstanding, in the vast majority of American cities and towns, there is only one major newspaper, precisely because it is so expensive to start one. The Encarta online encyclopedia states: “By 2000 only about a dozen cities in the United States had separately owned competing newspapers, and in 2002 Canada had only eight cities with competing newspapers under different ownership.”

Posner then shifted gears, applying his case about declining barriers to entry to television networks. “The current tendency to political polarization in news reporting is thus a consequence of changes not in underlying political opinions but in costs, specifically the falling costs of new entrants,” he wrote of television news. But starting a television network is the most expensive enterprise in the media world; doing so can cost tens of millions, perhaps hundreds of millions of dollars. For example, when News Corp. chairman Rupert Murdoch founded Fox News Channel, the start-up costs were estimated at $500 million. There are likely many ideologically motivated people who would love to start their own news channel, but the economic barrier to entry here is extraordinarily high.

Posner then made this extraordinary claim: “The rise of the conservative Fox News Channel caused CNN to shift to the left. CNN was going to lose many of its conservative viewers to Fox anyway, so it made sense to increase its appeal to its remaining viewers by catering more assiduously to their political preferences.” What evidence does Posner offer for his contention that CNN has “shift[ed] to the left”? None whatsoever. In fact, since 2004, Media Matters for America has documented hundreds of instances in which CNN has passed conservative misinformation on to the public.

And Posner's contention that CNN has shifted left “to increase its appeal to its remaining viewers by catering more assiduously to their political preferences” made little sense either. According to a 2004 survey by the Pew Center, CNN's audience is 44 percent Democrat, 25 percent Republican, and 25 percent independent (6 percent of respondents said they belonged to some other political affiliation or didn't know). While Democrats may have a plurality, more than half of CNN's viewers are not Democrats. Another Pew poll in 2002 found the ideological makeup of Fox News and CNN viewers to be almost identical: Fox's audience was 46 percent conservative, 32 percent moderate, and 18 percent liberal, while CNN's audience was 40 percent conservative, 38 percent moderate, and 16 percent liberal. So it would be an odd business decision for CNN to “shift to the left” to please its audience.

Posner showed where he is coming from when he cites the conservative hit parade of media bias claims. “The bias in some of the reporting in the liberal media, acknowledged by [former New York Times public editor Daniel] Okrent, is well documented by [William] McGowan [author of Coloring the News: How Crusading for Diversity Has Corrupted American Journalism (Encounter Books, 2001)], as well as by Bernard Goldberg in 'Bias' and L. Brent Bozell III in 'Weapons of Mass Distortion.' " This may be the first time anyone has referred to Goldberg's Bias as “well-documented.” In fact, the book contains no documentation of any kind and is full of claims that are either unsubstantiated or factually incorrect (for a debunking of Goldberg's claims, see this article by Geoffrey Nunberg in The American Prospect).

Posner also claimed: “The conventional media filter out extreme views to avoid offending readers, viewers and advertisers; most bloggers have no such inhibition. The argument for filtering is an argument for censorship. (That it is made by liberals is evidence that everyone secretly favors censorship of the opinions he fears.)” Which “liberals” was Posner talking about? We have no idea, since he didn't say. And Posner seems to believe that if you think editing is sometimes a good thing, then you “favor censorship.” He seems not to consider the possibility that one could simultaneously oppose censorship and think it's a good idea to edit both blogs and news products.

Posner may be correct that the proliferation of ideological blogs and websites has created a media environment that contains more polarized rhetoric than in the past -- although there has always been plenty of ideological rhetoric around (just listen to talk radio). But his contention that the products of large media companies have also moved to the ideological edges is difficult to substantiate; indeed, he didn't even try.