Who: People in the News

More than a decade ago, studies first noted that people were disappearing from the front page (see Danielson & Lasorsa, 1997). When feminists took measurements to see how women were faring, they expected to find female references increasing modestly while males continued to dominate front pages. The New York Times and the Los Angeles Times did refer more often to men than to women, but references to both genders had been declining over time. A follow-up analysis added other terms related to the human element in the newspapers, such as wife, father, child, or in-law. It confirmed that references to gender, such as he and she, had dropped, and the broader measure of human relationships had also been declining for decades. References to the human element have fallen by more than a quarter since 1900 (from more than 750 to fewer than 500 references per 10,000 words in 1960, with little recovery since then, Danielson & Lasorsa, 1997). The newspapers were presenting society "not from the point of view of single persons but from the point of view of groups" (p. 109). The newspapers also increased their use of quantitative words (from 140 in 1900 to just under 260 by the 1970s), leading the authors to wonder: Do these trends reveal that "people as individuals are becoming less important and numbers more important" (p. 109–10)?

Are individuals less important? To find out, a colleague and I started out by counting how many distinct individuals and groups appeared per story in content throughout the newspaper (Barnhurst & Mutz, 1997). The results suggest three observations (Figure 2.1).

Individuals who are identified only by name became a minor element. Just about everybody in the newspaper these days is described, usually by a group affiliation. And the number of groups in an average news article went up more than half by the end of the century. My follow-up study that looked at the same newspapers in 2001, after they had established Internet editions, found that these trends went further (Barnhurst, forthcoming). Named individuals have all but vanished, described individuals jumped, and groups continued the upward trend that had paused in the 1970s. Here are some examples from the studies.

On April 16, 1894, the New York Times reported a railroad accident in Hazelton, Pennsylvania. A front-page story listed every casualty by name, occupation, and medical condition. The format emphasized individuals:

Likewise, an 1894 Chicago Tribune story about an accident at a school listed each injured child’s home address, age, and medical condition in a separate paragraph (April 10, pp. 1, 7). In 1914, individuals’ names still ran in the lead paragraph in an article describing an accident:

By 1954, however, there was a greater emphasis on groups rather than individuals. Here is a typical lead: "A veteran West Seattle high school teacher who refused to say whether she ever belonged to the Communist party will be dropped from the Seattle school system, the school board announced Tuesday" (Portland Oregonian, April 21, p. 1). This job-related story has four groups — a high school, the Communist Party, the school system, and the school board — surrounding a person who is identified only by a job title that connects her to another group, a high school (the reference to gender also links her to other women).

In contrast to the earlier train and school accident stories, a typical 1994 report on a jetliner crash focused on the national origins of those involved and listed no names of casualties (New York Times, September 28, p. A9). In news stories today, most individuals no longer stand as individuals per se. The naming of individuals exists only in monuments, the Vietnam Memorial being the best known, and again the group connection tends to overpower the individual name. In news, names instead take on increasing importance as exemplars of particular human categories — Russians, women, or Latinos, say — or of particular types, such as workers, athletes, or small businessmen.

Citing someone’s group affiliation is a balancing act. In a well-known letter to the Friar’s Club in Hollywood, Groucho Marx wrote, "Please accept my resignation. I don’t care to belong to any club that will have me as a member" (cited in Andrews et al, 1996). The comic actor plays on this irony: Although his words seem self- deprecating, they also assert his own fame. People who matter don’t depend on group affiliation.

The push and pull between naming and grouping has played out in the news throughout U.S. history. In the colonial era, government officials were introduced with their full names. An edition of the Pennsylvania Journal in 1757, for instance, identified the New Jersey governor as "his Excellency Jonathan Belcher," when introducing his address to the General Assembly (Barnhurst & Nerone, 2001, p. 32–3). A less potent figure would be named less thoroughly. Reporting about a storm at sea in 1767, for example, the paper says, "Capt. Greenway . . . informs us . . . that one of his people was washed overboard and drowned, and himself narrowly escaped" (p. 42). The title and surname separate the captain from the unnamed sailor. The Connecticut Courant that year contained this item about a fire: "A few Days ago, a Dwelling-House, belonging to one Waterous of Marlboro’ was consumed by Fire, together with a great Part of the Furniture." The property owner has a surname, but is introduced as "one Waterous," to indicate his limited claim on public attention. Another item from the same issue identifies a woman, who lost quadruplets after delivery, only as "the Wife of one Flint of Windham" (p. 41).

These examples illustrate the power of naming. It separated out a class of historical actors. They had names connected to property deeds in public records, and they held the right to vote as citizens interested in public life. The historical classes were capable not only of taking action but of saying what it meant. As verbal actors (Ekecrantz, 1997), they could stand on their own words. No one had to speak for them, and so their group affiliation could go unstated. Ordinary people were mere physical actors, who left no mark on ideas or history. Colonial news stories about "a farmer’s wife" or "a laborer" or "a Negro servant" clearly showed who didn’t matter as an individual. The class of everyday people engaged in labor, and their work followed regular cycles that reproduced conditions of the material world without controlling it consciously. Instead they were subject to control, like any other property. In news, they were reduced to social categories. Group identity put them in their place.

One analysis of who appeared in news coverage looked at an event that occurred with surprising regularity for a century after Lincoln: the death of U.S. presidents in office. We found that important people were named and, increasingly, pictured, not only in the elite New York Times but also in the popular Chicago Daily News (Barnhurst & Nerone, 2001, see Chapter 5). Ordinary people were depicted as an undifferentiated mass of mourners along the routes of trains that moved the deceased president to his grave. When William McKinley died, the Times depicted the category of workers as an elemental flow from nature:

As ever more subdivided groups appeared, stand-alone individuals began to vanish. In 1894, nearly half of all individuals in the news went by name alone. The front page of the New York Times contains a crime story about two Long Island residents who "were speeding their trotters on the Oyster Bay Turnpike Saturday evening. As they tried to pass, their vehicles became locked together" (April 16, 1894). The drivers jump out, argue, throw punches and begin a "tussel." We see a gristmill and pond nearby, and suddenly the fight splashes into the canal. As the loser is about to drown, a passer-by intervenes. The case goes "this morning" before a justice, who fines the aggressor fifty dollars. Each person can simply be named — not tagged as representing particular subgroups — because the story’s the thing. Narratives like this one (which ran about four column inches) packed front pages with people.

The practice all but disappeared over the course of the century. Journalists at first adopted one standard way to identify individuals, by name and domicile. In the late 1930s, a daily column called "City Locals" in the Beverly Evening Times in Massachusetts contained items like these:

The reporter, a kid on summer vacation whose dad knew the publisher, would walk along the main commercial street of Beverly, a small town outside of Boston, "stopping in one store after another and asking the employees what was new in their lives. Any birthdays, new arrivals, deaths? Illnesses? Any visitors at home? Any vacations planned? Home alterations? Parties, retirements, confirmations? College acceptances? Dean’s lists?" (p. 37).

Name and address became the norm for ordinary people, rooting them in a neighborhood that labeled them with all the tacit knowledge about what kinds of people lived there (see, e.g., Ettema & Peer, 1996). Only notables and celebrities escaped the binds of social locations. They were identified by name alone. The sociologist C. Wright Mills called them "the Names that need no further identification" (1956, pp. 71–2).

As groups rose in importance, two things happened to ordinary people. Their street addresses dropped out of the news (who wants a home address published in an item about a pending trip that will leave it vacant?) The new standard identification used group affiliation. The other thing that happened was that fewer ordinary people were named at all. Instead news focused more on the well connected and well to do. In the 1950s, sociologists found that newspapers named the elite far more often than anyone else, while also mentioning hundreds of voluntary associations (Janes, 1958).

It’s possible that the shift from individuals to groups, which peaked in the 1970s, reflected social reality, that people themselves took their identity increasingly from groups. By that decade, Vietnam War protests, the women’s movement, gay liberation, and many racial and ethnic groups had taken center stage in American social and political life. In popular culture, Alex Haley’s Roots was published in 1976, followed the next year by the mini-series on television. But earlier periods in U.S. history also had groups at center stage. It’s also possible that the shift to groups reflected the way that journalists viewed the world. Instead of rousing stories of individual action, reporters and editors wanted to show the complex interplay of social forces. That required groups.

Now, after the turn of the new century, individuals act as stand-ins for groups. A recent front page of the Sunday New York Times features two individuals (August 4, 2002): "Joshua Hastert, 27," is "the eldest son of J. Dennis Hastert, the speaker of the House," and one of "numerous relatives of Congressional and administration officials" who are "employed in lobbying shops around Washington" ("In Capitol, Last Names Link Some Leaders and Lobbyists," pp. 1, 17). In a foreign report, "Adrian Wilkinson, 50," is like "most of Zimbabwe’s white farmers" ("For Zimbabwe White Farmers, Time to Move On," pp. 1, 4). Both of them appear in pictures, giving the impression that the news focuses on people. But groups outnumber individuals on the front page by more than two to one (54 to 22, to be precise) and dominate the top half of page one. In column one, for example, the fourth paragraph contains only one individual (who was already named and described in the lead paragraph of the story) amid six groups arrayed for a legal battle:

None of this list of groups is specified — and no individual speaks for them — until after the story jumps from page one to fill an interior page. Today more groups stand alone as the actors in news, and ordinary, named individuals almost never do. Newspapers have left behind one of their most powerful attractions to most readers: " ’Tis pleasant, sure, to see one’s name in print," Lord Byron wrote in a couplet, "A book’s a book, although there’s nothing in ’t" (1905, line 51, p. 242). The substance of news was always secondary to its power to name people (and involve them by naming them). But people’s names have given way because individuals can no longer stand alone. They act primarily through groups.

Even so, institutional actors didn’t take center stage in the popular imagination for some time. In the nineteenth century, people still held individuals responsible for the actions of businesses, sometimes using violence to make their point. In the case of newspapers, it was not unusual for an editor who published offensive material to be tarred and feathered and then run out of town on a rail (Nerone, 1994).

When he worked for the Courier Journal in the late 1970s, Michel Marriott was one of very few black reporters in Louisville. He would get calls from people "who never called the newspaper before," asking for his help (Marriott, 2001). Often there was little he could do, other than listening and taking the caller seriously. But sometimes he managed to help someone hold an institution to account. He tells the story this way:

If more groups dominate the news, it follows that more people in the news play roles as leaders of those groups. The usual journalism is about those officials (Hall et al, 1978; Fishman, 1980; Ericson, Baranek & Chan, 1989). Research in the 1970s showed how news organizations cast their net to gather stories (Tuchman, 1978). Most people slip through the holes in the net, which is built to pick up stories from the people most likely to know about them: officials. But sociologists say the share hasn’t changed, at least in the past few decades.

In newspapers, an early study of sources reported that more than half were government officials (Sigal, 1973). In the New York Times and the Washington Post, front pages of the 1950s and 1960s were peopled by officials of American and foreign governments, in part because in any situation, reporters want to speak with someone in charge. A similar study conducted ten years later concluded that not much had changed (Brown et al, 1987). Although new technology intervened and the news industry reorganized, officials remained at the forefront in news stories, reinforcing the existing system of power and legitimacy.

For television, the sociologist Herbert Gans divided people on the evening news into two categories, knowns and unknowns, and reported that the knowns appeared about three times more often than unknowns (Gans, 1979). Among the widely known people in 1967, presidents and presidential candidates, members of Congress, and federal, state, and local officials amounted to about 45 percent of all people in the national news. Gans concluded that reporters relied on these knowns because they were not only accessible but also suitable. They were ready and willing to act as sources (if not in fact seeking attention), and they were close by as well, both physically and socially. The sources were suitable because they were connected to the centers of power, had access to information, and could be counted on to speak articulately. A study conducted in the early 1980s found no change (Whitney et al, 1989). There was about the same share of knowns and about the same share of political sources on network evening newscasts then, and the same was true in the early 1990s as well (Jones, n.d.)

Each of these studies examined sources over the short run, but taken together, the research looks at newspapers for about fifty years and network newscasts for almost twenty years. In both print and broadcast news, the focus on official sources stayed about the same, sociologists say. But my long-range research shows that in crime, accident, and employment stories, the roles people play in the news has had some significant shifts (Figure 2.2).

Studies that count words per page, rather than people per story, have measured a persistent growth in terms relating to officials on the front page, with steep increases since the 1960s (Danielson & Lasorsa, 1997). Regulatory terms related to the federal bureaucracy have also increased, as have terms referring to the national government in general. A related study found that verbs used to attribute information to officials have increased (Barker-Plummer, 1988).

What is the relationship between actors in the news and the officials in charge? In a typical example from the New York Times on the Web, an article on the guest worker program, "Calls for Change in Ancient Job of Sheepherding [sic]," begins with several paragraphs telling the story of a shepherd in the Mojave Desert. To close the anecdote, the writer says, "The man, who would not let his name be used for fear of angering his employer, is typical of shepherds in the United States, about 800 of whom toil in California." The article then moves on to groups, officials, and their affiliations: the unchanging trade of shepherding, the geographic origins of workers, the broader labor pool, the bodies regulating wages, the California legislature and its proposed program, and so on through lobby groups, industry groups, labor advocacy groups, and legal aid groups. Official sources of various sorts speak for half of the ten groups in the story; the rest stand alone. Individual actors without an official role do not appear again until the end of the long article. After a general paragraph describing shepherds at work in Fresno and Kern Counties in California, the story closes with the man from the opening anecdote. "I get depressed here very often," he says.

This powerful story turns the individual into a rhetorical device. Gans (1979) noted how television news reports most often used individuals as epitomizing cases. Reporters call the technique the anecdotal lead. These exemplars have become almost universal in print stories and appear in about half of television news stories (Zillmann & Brosius, 2000). Exemplars have a strong influence on the audience and may particularly sway opinion if the anecdote is emotional or sensational. The anecdotes can be misleading when journalists choose an atypical or dramatic exemplar. Even when journalists explain that the exemplar is unusual and give details about the low frequency of similar cases, members of the audience will remember the anecdote and forget the statistical rate. If for example a story says two thirds of small family farms operate at a profit, but then tells stories about failing farmers and bank foreclosures, people remember the failures, not the success rate. Using exemplars is common in trend stories, especially if the trend being reported involves something threatening (Berger, 2001). In the story about shepherding, it’s clear that the vignettes describing an individual are meant to build a dramatic opening and closing for the real work of the report, to cover the official exercise of power.

A news report, like biography, "assumes the value of an individual," in the words of novelist A. S. Byatt (2000, p. 26). "Whether you see that individual as unique or as a type depends on your view of the world." Journalism justifies itself and its civic mission by aiming to respect people as unique, but the everyday content of news treats the individual as a type. Individuals stand for problems that only officials can solve.

Besides playing the role of Robin Hood, journalists take on another role, as a watchdog. The press keeps an eye on the public estates (the government) and alerts its master (the voters) of any disturbance. For example, a reporter working on one of the zoned editions of the Pittsburgh Press gets a tip and files a story about a suburban town that forgot to pay withholding taxes for employees. At the next town meeting, citizens complain about the oversight, and the town council expresses outrage that the newspaper published a story. But people say, "No, then we wouldn’t have found out." The town has to raise taxes, but there are also consequences for the council members: they get voted out in the next election. It’s gratifying for the reporter to see people rally around the newspaper, and to see the newspaper make lives better by correcting errors in government. "It’s a small story," the reporter says, "but it’s important to the people involved" (Anonymous, 2001).

Work that costs officials their jobs also burns a reporter’s bridges (Ericson, Baranek & Chan, 1989). Too much of it would put a newspaper out of business. Of course, in both of these cases, the burnt officials were then replaced, but any future leader would be wary, making it more difficult for the newspaper to cover routine actions of the town council or the school system. If journalists want to get the news, they have to make peace with official power.

A century ago, like many working-class people, journalists had little use for experts. News was a commonsense affair, written by an everyday observer, and targeted to an ordinary reader. Reporters continued to be generalists well past mid-century. Joseph Lelyveld, who covered South Africa in the 1960s before his rise to editorial positions at the New York Times, said that as a reporter he lived by the maxim, "Every person is an expert on the circumstances of his life" (1985, p. 113). The other kind of expert, an observer who watched events from a distance and had no direct stake in them, was the object of scorn in memoirs by journalists for much of the century. Sociologists also observed that attitude as late as the 1970s, when they studied journalism by spending time with reporters and editors in the newsroom. Journalists viewed Gans himself with suspicion as an academic (1979).

Somehow, despite the disdain of previous decades, experts recently became a fixture in the news. In the 1980s, researchers began to notice the large numbers of expert sources. Some turned up in obvious places, such as the health care specialists who were the sources most frequently mentioned and most often quoted in medical news (Stempel & Cuthbertson, 1984). Others turned up in stories that otherwise focused on ordinary people, such as the coverage of victims caught up in a 1985 hijacking of a TWA flight (Atwater & Green, 1988). On the three national television networks, experts, former government officials, and others were the most frequent sources after the hostages and their relatives themselves. When all sorts of news stories were studied, experts such as academics, professionals, and former officials commenting on their specialty areas accounted for around 20 percent of all network news sources (Whitney et al, 1989).

Long-term research shows that the number of experts has been growing for some time, not only in census data (Brint, 1994), but also in the news. The frequency of expert and professional words appearing on the front pages of the New York Times and Los Angeles Times roughly doubled, from around twenty in the 1880s to around forty a century later (per thousand words, Danielson & Lasorsa, 1997).

When we measured the share of sources in the average newspaper story, we found that since the late nineteenth century the number of outside sources grew gradually until around 1974, and then increased more rapidly into the 1990s (see Figure 2.2).

Experts multiplied on television. People whom the newscasts identify as analysts, consultants, academics, former officials, and the like are what media scholar Lawrence Soley calls news shapers (1992). In the decade from the late 1970s to the late 1980s, the number of news shapers almost tripled on the three networks. By 1988 there were three in the average newscast. Appearing on national television enhances their expertise. Through what sociologists call status conferral, experts acquire an aura of knowing, just by appearing in the media. Journalists intensify that aura by relying on the same experts repeatedly. Soley found that thirty-four shapers accounted for nearly half of all appearances, and two shapers appeared as many as ten times. The rule in news is, the more you’re cited, the more journalists will cite you. Some of the most widely used experts in the news, Soley shows, have very little expertise beyond an ability to deliver dramatic, quotable sound bites.

During the Gulf conflict in the early 1990s, researchers found that the large number of experts created an illusion of depth. Rather than understanding the Gulf region in depth, based on knowing the language, history or culture, shapers supply a different kind of expertise. Theirs is based on knowing the players, having an inside scoop on policies, and being willing to make predictions. TV experts deliver these three p’s — players, policies, and predictions — using a fourth, patter, the ability to talk in succinct sound bites (Steele, 1995). Other research found the same notion of expertise at work in major U.S. newspapers (Hallin, Manhoff & Weddle, 1993).

Journalism embraced this kind of experts, and people paid attention to them. When experts and others who comment on network news shift positions on major issues, research has shown, a parallel shift occurs in U.S. public opinion (Page, Shapiro & Dempsey, 1987). For example, during Senate consideration of the second Strategic Arms Limitation Treaty (SALT II), each network put retired generals and arms experts on the air. They argued against the treaty because it had verification difficulties and would tip the balance of forces in favor of the Soviet Union. The research showed that public support for the treaty then dropped by almost a quarter. Such changes occurred for other policy issues over the fifteen years included in the study. Television journalists rightly conclude that expert sources do have an impact on the public.

As the cult of expertise grew, journalists themselves began joining the ranks of news shapers, acting as expert sources on topics they cover. In newspapers, they tend to turn up as unnamed sources. A reporter from a major newspaper, who spoke (but not for attribution) at the Media Studies Center at Columbia University when I was a fellow there in 1992, described using himself as a source when covering a country in Asia. When he first began work as a correspondent there more than a decade earlier, he relied on sources called Asia-watchers. After he had been in the country for a few years, he thought he understood what was happening better than most of the Asia-watchers. From that conclusion, he took the next step: at times he explained events by attributing his own ideas to an unidentified observer. The practice was part of the growing use of what are called veiled sources, which the American Newspaper Publishers Association began examining in the 1970s (Brown et al, 1987).

In other news media, journalists appear openly as expert sources. Newspaper reporters go on television and radio to talk about events they are covering, and news anchors and hosts in many cases treat reporters from their own networks as experts by conducting interviews with them on the air. Content analyses of news indicate that the phenomenon has generally increased in the past quarter century. The well-known study of television sound bites from 1968 to 1988 observed a large growth in what are called wrap comments (Hallin, 1992), the concluding words that summarize the correspondent’s expert conclusions. My own detailed study of journalist speech on National Public Radio also found more journalists acting as experts (Figure 2.3).

The share of speech by experts on NPR daily news programs grew by half since 1980, the first presidential election year when both "Morning Edition" and "All Things Considered" were on the air. The appearances of expert sources dropped off during the 1996 election year, when Senator Bob Dole mounted a lackluster challenge to the incumbent President Bill Clinton. The share of journalist appearances as expert sources was small at first, but grew swiftly over the next four years. After declining in 1992 and 1996, the share of speech by journalists as experts ended up four times higher than it had been twenty years earlier (Barnhurst, 2003a).

Besides the share of speech, the length of speech for journalist-experts grew. All the voices on the NPR reports saw an initial jump in average speaking length between 1980 and 1984, but the voices of journalist-experts increased the most of any group. Other sound bites then grew shorter in subsequent years, but journalist- experts, after an initial decline, grew again, reaching almost a minute per speech by 1996. Although the length of their speech then dropped in 2000, the pattern was clear: journalists as expert sources spoke longer on average than all other journalists (and all other speakers) every year after 1980.

We found a similar pattern on television news (Steele & Barnhurst, 1996). In 1968 correspondents asked questions of sources. Anchors didn’t ask questions at all. Twenty years later the opposite became more common. Anchors asked more questions and correspondents gave more answers. The anchors had begun treating correspondents as experts. For example, in a 1988 ABC News broadcast, Peter Jennings gave this lead-in to a report:

In his revised role, Jennings treated the correspondent as a source. Hallin’s sound-bite study accounted for part of this shift by noting an increase in what he called "truth squad stories," in which journalists played the part of experts who corrected the claims of candidates and their advertising (1992, p. 19).Being answer-providers gave correspondents higher status by recasting them as specialists. In other words, McWethy was no longer just any correspondent but a specialist in national security.

It didn’t have to work that way. When Michel Marriott worked for the Marion paper, he was assigned to cover the annual event. But he thought, "Man, the real story is the journey." He asked his editors to let him make the entire walk, and because he was a young reporter on his first full-time job, the editors said yes. He did the whole trip, from spending the night in a sleeping bag at the scoutmaster’s house, to the end of the walk. "It was tough," he told me. "It was a tough thing to do." He got to know the kids well and became a buddy, not just an outsider. He filled four reporter’s notebooks, front and back, with stories about real people, the ones who faltered and the ones who helped them. He had so much stuff when he got back that he wrote not only the annual story but also a cover article for the Sunday magazine. "It was a big deal," he said (Marriott, 2001). His story gave ordinary individuals a greater voice. They didn’t have to have status in the official hierarchy of the troop. They didn’t have to be experts. But his reporting took more time and his writing ran longer.

Have news organizations turned to expert and official sources because it’s faster and cheaper? The two main uses of a journalist’s productive time are reporting (research, groundwork, interviews) and writing. Longer stories take longer to write, no matter whom the sources are. Because stories in general have been getting longer for more than a century, the writing itself didn’t change the choice of sources for a story. What about the time spent doing the reporting? When news had more individual actors and fewer experts and officials, someone was spending time with the un-titled people, or at least paying more attention to gathering events at the grass roots. Today, reporters are still talking to people — their anecdotal leads come from that work. In the example from the New York Times on the Web, the reporter traveled around the Mojave Desert, meeting and interviewing many more shepherds than the one in the published story. The cost of doing that sort of reporting didn’t shift attention to officials and experts. Reporters write fewer stories today than they did a century ago, and they spend more time on both reporting and writing. Research comparing sources to staff bylines shows that individual reporters have some latitude when deciding whom to interview and include (Hallin, Manoff & Weddle, 1994). They slowly adopted the practice of focusing attention on powerful, expert sources for other reasons, not to save money. Stuart Hall called this over-accessing the powerful and privileged (Hall et al, 1978). There is no inherent reason that news could not have kept the same ratio of individual actors to authoritative sources as it grew longer, but it didn’t. Journalists chose the authorities. Why is that so?

Ben Bradlee joined the management of the Washington Post in 1965. Soon he went on a hiring spree and shook up the existing staff. One of them, Gerald Grant, recalls:

Once upon a time, the scion of an old Boston family grew up and went to Harvard, but then the United States entered World War II. He left school and joined the Navy, becoming an ensign and serving in the Pacific. When the war was over, he moved to Washington, D.C., where he eventually became prominent and powerful. Does this career sound familiar? It describes not only John F. Kennedy but also Ben Bradlee himself (1995), and it’s more than a coincidence. Not long after starting out as a journalist, Bradlee and his wife were taking a walk pushing a baby carriage when they met a Georgetown neighbor, the junior senator from Massachusetts and his wife Jacqueline, who were also pushing a baby carriage. The two families immediately became good friends, eventually even vacationing together. Bradlee wasn’t upwardly mobile in the sense of moving from one social class to another, as his $100,000 trust fund attested (he was fifty-second in the line of Bradlees who went to Harvard), but his reputation grew as John F. Kennedy became more powerful.

People in any line of work move up when they make social connections in high places, and their status also grows when they succeed in attracting higher class members to begin with. The stature of journalists, as documented in Washington since the 1930s, has risen since the end of World War II. Daniel Patrick Moynihan, writing in 1971, said that "twenty years and more ago the preponderance of ‘the working press’ (as it liked to call itself) was surprisingly close in origins to working people generally. They were not Ivy Leaguers" (p. 43). In the 1930s, about half of Washington journalists held a college degree (Rosten, 1937), but that figure grew to more than 80 percent by the 1960s (Rivers, 1962). By the end of the 1970s, more than 90 percent of Washington reporters were college educated, compared to about 17 percent of the general public (Hess, 1981). About 35 percent of the reporters had highly selective alma maters: Ivy League schools, small private colleges like Swarthmore and Carleton, and prestigious universities like Stanford and the University of Chicago.

Reporters outside of Washington also became more highly educated. The share of college graduate-journalists nationwide grew from 60 percent in 1971, to 75 percent in 1982, and 82 percent in 1992 (Weaver, 1999). The share of journalists holding graduate degrees also increased, especially among the Washington press corps. Most U.S. journalists are white and male. Women account for less than one third of U.S. reporters. Non-whites are also underrepresented. Reporters from racial and ethnic groups grew from 4 to 8 percent between 1982 and 1992, lagging far behind the job growth of the same groups in other sectors. In the Washington press corps, the share of minorities and women is even smaller. Census figures indicate that these minority groups make up about 24 percent of the U.S. population.

Reporters, especially at elite news outlets, are mostly well-educated white men. As a group, they match the description of what sociologists call the power elite: upper income, professionally employed, urban sons of comfortable families. As members of a key institution — the media, where "serious public figures too, must now compete for attention and acclaim" — they share the origins, education, and styles of life that make it easy to intermingle with other power elites (Mills, 1956, p. 75).Research on the press and corporations, trade associations, policy groups, civic groups, and social clubs reveals "a web of affiliations" linking newspaper-owning companies "to the nation’s power structure" (Dreier, 1982, p. 298). Similar connections exist with the power structure at the local level (Soloski, 1989). Does that intermingling matter?

It does to journalists, of course. Those who are well educated, connected, or fortunate enough to get jobs in journalism have seen their collective fortunes rising. Many of them still put in long hours for low pay with little chance of advancement, especially at small town newspapers and within chain-owned operations (Hardt,1998). This is typical of many professional groups. Those who work within large, prosperous organizations can afford to adopt lofty ideals, compared to those who work in smaller concerns at the mercy of market forces (Brint, 1994). Journalism has come late to professionalism, and while other groups have moved away from their early aristocratic noblesse oblige and now base their authority on expert knowledge, journalists have only recently moved in that direction. In any case, the lot of reporters generally was much worse fifty or one hundred years ago and has improved along with that of other social and cultural professions.

By the late 1960s and mid-1970s, when Gans conducted his ethnographic studies among journalists, he found that "they learn from peers and personal contacts, notably relatives and friends." As a result, they "move within a relatively small and narrow aggregate of sources," "dominated by the people they contact or who contact them regularly" (1979, p. 126). Journalists considered their work companions, families, and neighbors and the people they met at parties credible representatives of the audience. They also used themselves as sources, relying on their own lives and experiences for story ideas.

Subsequent studies among newspaper and television journalists confirm what Gans found (Hess, 1981; Sumpter, 2000): journalists come up with story ideas from people like themselves, they rely on the same sources repeatedly when reporting and interviewing, and this small circle of contacts stands in for the audience while journalists are writing stories and choosing how to play them in print or on the air. The variety of contacts also tends to contract as journalists move higher in the hierarchy of news organizations. Reporters on the job get to know a greater variety of people than do editors who work all day in the newsroom.

Reporters and editors imagine their audience as being much like themselves or like what ethnographers call the proximate audience made up of people they know. News has always been personal. In the colonial era, editors relied on personal contacts with gentlemen and other editors, and that hardly changed despite two centuries of technological transformation. News is still "an information system based on personal contacts" (Hess, 1981, p. 126).

Personal patronage was once the norm for government. A new president, governor, or mayor in the nineteenth century could fill many official jobs by appointing friends, acquaintances, friends of friends, and perhaps even family members. The patronage system was one of the chief targets of the Progressive Movement, which helped pass laws establishing a professional corps of civil servants who win their jobs through merit and hold them even when administrations change. Only the upper echelons of managers continue to be appointees dependent on political bosses. A similar movement away from dependence on charismatic leaders and toward rationalized systems has occurred throughout society (Weber, 1958). Large corporations and even religious groups, although not obligated by law, now customarily hire and fire workers at the lower levels based on something like a merit system. In the study of politics and organizational behavior, systems based on personal patronage are seen as unjust or worse, corrupt. They rely on what is called personalism, which focuses power on an individual leader who is in a position to pass out favors.

The other face of personalistic systems is called clientelism, the dependency of any group or individual who owes personal fealty or is in the leader’s debt. Clientelism is more complicated than that definition suggests, because the agencies of power also become the captives of their dependent clients. Clientelism is a key element in business and political life in parts of southern Europe and Latin America, and also characterizes media organizations (Hallin & Papathanassopoulos, 2002). Even in the United States, an interlocking system of direct personal influence still drives news businesses, newsgathering, and journalists’ concepts of the audience.

When sociologists and anthropologists do field work in the newsroom, they find the relationship of journalists with the audience oddly paradoxical. Reporters and especially editors have a hunger to know who is out there reading and viewing, but resist knowing the answer. They rejectal most all of the available information, except what they can gather from a few proximate stand-ins they know personally. Gans found that journalists "have trouble crossing social barriers that separate them from strangers" (1979, p. 126). Strangers, mostly from other classes, ethnicities, or gender identities, and often with less education, have difficulty reaching journalists as well. A case study of who gained access to and attention in a Boston newspaper found that non-elites had a hard time because they don’t speak standard English (Goldenberg, 1975). Reporters want quotable speech that’s dramatic, articulate, and concise. News now requires the language of elites, which poor people, immigrants, and outsiders don’t normally produce. To get beyond the usual sources, one newspaper assigned reporters once a month to eat lunch with someone they didn’t know and hadn’t used as a source, and then report on their conversation to the metro editor (Sumpter, 2000).

Gans observed that journalists fear the massive numbers of their audiences. These large aggregates seem to exercise poor judgment by preferring gossip to real news. Social science research has tended to confirm that view. Techniques for large-scale surveys of public opinion began emerging in the mid–twentieth century (Herbst, 1993), and the results seemed to leave little doubt about the low levels of information that most Americans possessed. The average U.S. citizen could not identify important political leaders and did not understand the processes of government or the workings of other institutions and corporations. That hasn’t changed. If journalists of the nineteenth century considered the public attentive and astute, this new evidence from social scientists did not sustain the same respect.

The other form of feedback available to journalists comes unsolicited from individual readers, viewers, or listeners. Reporters and editors consider audience mail highly idiosyncratic. They treat the letters, phone calls, and electronic mail they receive just the way anyone treats messages or calls from strangers: with suspicion. And reporters are right. Gans found that letters to NBC News came from an entirely unrepresentative subset of the viewers. When call-in radio programs became popular in the 1990s, some editors began listening in, trying to learn more about the public out there (Sumpter, 2000). They gave it up after discovering what they already knew from their earlier experiences with letter writers: they are older and more conservative, and they are mostly disgruntled — unless they are fans of someone on the staff and want an autographed picture, something that happens mostly at network programs. Television reporters who answered calls from viewers routinely threw their notes in the trash, except during some media events, such as the verdicts that set off rioting and protests after the Rodney King beating in Los Angeles, when they receive large numbers of unexpected calls from people like themselves (Jacobs, 1996). When selecting what to publish in the letters-to-the-editor column, newspapers follow the same patterns set for other sources, preferring officials, experts, and other "authorized knowers" (Ericson, Baranek & Chan, 1989, p. 362).

A typical white, male, middle-class journalist is stuck with the people he knows and can rely on: people who are like him socially, psychologically, economically, and so on, sources he talks to regularly enough to know that what they say will pan out. When those people aren’t available, he goes first to people whose job titles and reputations make them seem most likely to be reliable. Studies have shown that who gets into the news depends on how closely they are connected to the news organization (Martin, 1988). Newspapers in Indiana and Kentucky published more stories with a greater number and wider variety of sources when a news staff member was directly involved in the community being covered. The diversity of sources is also greater for what’s called enterprise reporting, when journalists explore or investigate new stories (Hansen, 1991). But routine daily news produces a less diverse array of sources, with journalists and authorities talking among themselves (Seymour-Ure, 1974), in what might be called horizontal communication.

The resulting product, based on the personalism and clientelism of everyday news work and the narrow range of views those contacts express, is still what most citizens rely on to get information. Perhaps as a result, audiences see the news media as biased and unrepresentative (Dautrich & Hartley, 1999). Surveys of confidence in the press have shown steep declines over the period when the status of journalists among elites has risen. In the 1970s, around 25 percent of the public expressed a "great deal of confidence" in the press, but by the 1990s the share had fallen to around 10 percent. Research on audiences and newspaper content shows that newspapers and television newscasts in fact provide the greatest exposure most U.S. citizens get to views unlike their own (Mutz & Martin, 2001). They encounter very little divergent opinion at work, and almost none in private activities within voluntary associations and at home.

But the public is also stuck. The press is the only game in town, the only place they can easily go to find out about other people, important elites, yes, but also ordinary people with experiences and views unlike their own. Alternative outlets do exist and produce oppositional news that follows the conventions of mainstream journalism, but most U.S. audiences have no access to anything like Pacifica Radio (Eliasoph, 1988). The Internet at first seemed to promise a wider array of perspectives on the news, but that hope quickly faded. My own attempt to listen to Pacifica Radio over the Internet proved impractical because of obstacles ranging from software limitations to differences in time zones. Research on Web-based news shows a growing concentration of traffic at just a few sites owned by large old-media corporations. Newspapers in many markets have moved aggressively to close down or buy up on-line competitors, reinforcing their monopoly control of local news and other information (Barnhurst, 2002).

A recent example can illustrate the interwoven strands that make up the who of news. In the 1992 presidential campaign, Bill Clinton began to give speeches and hold fundraising events among gay men and lesbians. No national politician had ever courted these voters so openly before. Gay men had begun to people the news regularly in the 1980s in health-related stories, but political leaders during the Reagan and Bush administrations preferred not to mention gay folks at all. Clinton’s action suddenly brought the queer vote into mainstream politics.

In the decade from 1992 to 2002 news organizations inadvertently ran something rare, a natural experiment in what happens to people when they become a regular element in the news. To measure the changes in who spoke on the air as gay groups moved out from the margins, I did a case study of National Public Radio (Barnhurst, 2003b).

From 1992 to 2000, reporters shifted away from using sources who were citizens involved directly in the events being covered. In a 1992 story on the anti-gay-rights amendment in Colorado ("Morning Edition," Wednesday, November 11), reporter Mark Roberts included two sound bites from involved citizens, both of them participants in a rally on Capitol Hill in Denver. Unidentified Man No. 1, in a talk at the rally, described the expressions of fear and support pouring in from LGBT communities around the country after the amendment passed. Unidentified Man No. 2, who participated in the rally, gave a reaction sound-bite: "It is a terrible feeling to be hated by other people for something that you had no choice in." Reporters referred to non-elite speakers as activists, and these gay men (most often unidentified) generally contributed coherent, fully formed thoughts in complete sentences. In other words, individual citizens who took courage and stepped forward were allowed to speak out for themselves. But ordinary citizens do not have office hours or job titles, they are hard to find on deadline, and they cannot be relied on to give good sound bites. They don’t move in the same social circles as most journalists. They may add a little spice and realism to reporting, but they lack authority.

Over the course of the next two elections, NPR news replaced active citizens as sources. Two types of speakers took their place. Uninvolved citizens almost quadrupled in number. The average report allowed one or perhaps two onlookers to speak in 1992, but included more than twice as many in 2000. The typical quote from non-elites became very brief, such as the repeated shouts of "That’s right!" at a rally against the civil unions law in Vermont ("All Things Considered," Tuesday, October 24, 2000).

In a nutshell, a habitual NPR listener heard the following sources in a typical week of gay-related coverage during the 1992 campaign: two citizens and one onlooker, one official and no spokespersons for groups. Two decades after gays and lesbians first appeared in mainstream political news, the habitual NPR listener might hear two officials and as many as three group leaders along with an expert, but only one active citizen spoke, along with several bystanders who uttered short phrases. The voices of experts and officials replaced the citizenry, and ordinary people’s voices were reduced to mere fragments.

The other reason news workers cannot gain a knowledge of the audience is conceptual. Knowing their public would, by definition, require that journalists yield control over their craft. Accepting the available information, or any conceivable information, would involve ceding news judgment to others. No journalist can do that and remain a journalist, especially given the anxiety journalists feel over their professional status (Zelizer, 1992). News judgment — the nose for news — is a key professional attribute for any reporter, the skill that gives pride of place to editors. It is not taught but learned through tough experience, through failure, through watching seasoned correspondents, and through hearing the stories of veteran editors. Journalists consider it an instinct (Sumpter, 2000). Obeying statistics would abdicate control to sales people, marketing types, advertisers, the business office, or someone else who doesn’t understand journalism. Obeying what comes over the transom would abandon control to disgruntled cranks. But there’s nothing else available from audiences. News people simply cannot know their public. That is their tragic flaw in this drama.

So like an unrequited lover, the journalist turns to surrogates. Sources play the third role in the triangle. Politicians and officials from every sort of group sell themselves to journalists. They want the journalist’s attention to gain the hard currency of news: public attention (the people again, imagined but out of reach). Journalists may satisfy their professional needs through their relations with officialdom, but they treat their sources with disrespect, cutting their sound bites shorter and disdaining their eagerness to please (Levy, 1981).

In their defense, it must be added that journalism is like any other profession. No one has access to the real "people." Studies of professionalism show that doctors, lawyers, and even university professors all claim to have a public mission. As professionals, we carry in our heads a picture of the public that is so deeply ingrained we hardly think about it. Theorists call this image an internalization (Robbins, 1993).When I asked Umberto Eco, a prominent theorist who is also a practicing journalist and author, about how he conceives of his audiences, he said they are like a ghost, but much smarter than editors and publishers give them credit for (Eco, 2002). Professionals argue that each decision we make serves our public, when in fact what we decide serves only what we imagine as the "people." The distinction is important. Because we don’t examine our imagined public very often or with very much rigor, we all fall into the habit of equating the people with ourselves. We feign to speak from two minds, but we are of only one. When journalists argue among themselves, one — usually an editor — will seize the high ground by claiming that her position serves the reader, listener, or viewer. As the sociologist Pierre Bourdieu observed (1990), the "people" is often what is at stake when elites struggle among themselves. Journalism has always covered elites, but today journalists themselves are among the elite.