In the academic community, Kathleen Hall Jamieson (1992) and Thomas E. Patterson (1993, 1996) argued that the press had lost sight of the challenges of governance, and had become consumed by political gamesmanship - by the daily pulse of tracking polls and the machinations and power plays within campaign organizations. The result, these scholars charged, was an inherently destructive style of campaign coverage that focused on the negative aspects of politics and drove citizens away from the political system.

Several well-known journalists, including David Broder (1991), James Fallows (1996), and Ellen Hume (1996), agreed that a great deal of election coverage was shallow and preoccupied with tactics. They argued that professional norms of objectivity and detachment had inflated to the point that many reporters and editors felt no responsibility for what happened in campaigns - no responsibility for whether politicians presented meaningful choices, or for whether citizens felt they had sufficient cause and information to participate. Said Broder: "We forgot our obligation as journalists to help [citizens] cope with this mass of political propaganda coming their way" (quoted in Fallows 1996: 253).

Out of these concerns, scholars and journalists developed a different approach to the coverage of politics - an approach that came to be called "public journalism" (Merritt 1995; Charity 1995; Rosen 1996). Instead of concentrating on candidates and consultants, public journalism proposed to shift the focus to issues and citizens - to cover public affairs from the public's point of view. From this changed perspective, public journalism - in the words of Jay Rosen, a professor at New York University who has been the concept's leading theoretical figure - would "uncover what concerns people in their role as citizens, and work to make politics and journalism address those troubles" (1996: 70). Journalists would listen carefully to citizens' concerns, and then prepare stories describing what the candidates would do about the issues that citizens raised.

The call for "public-style" press coverage swiftly attracted the interest of several prominent philanthropic institutions, including the Pew Charitable Trusts, the Kettering Foundation, the Twentieth Century Fund, and the Poynter Institute for Media Studies. Initial efforts at covering campaigns in a more public style - by the Wichita Eagle in 1990 and the Charlotte Observer in 1992 - drew substantial attention within the nation's newsrooms. By 1996, the Project for Public Life and the Press at New York University reported that about 400 public journalism projects had been launched by newspapers and broadcast outlets across the country (Conte 1996: 832).

But the public journalism "movement" - if it could properly be called that - also generated backlash from other newsrooms and scholars. These observers felt public journalists were taking on responsibilities that properly belonged to the political system, and thereby jeopardized the press's cherished position as an independent check and balance on government (Woo 1995; Raines 1996; Stepp 1996). There was also criticism from some in political circles, who charged that by giving preferential attention to the issues identified by voters, news organizations which practiced public journalism made it difficult for candidates to raise other issues that they believed were vital to the campaign (Jackson 1996, Kelly 1996).

Supporters of public journalism contend that their new style of coverage is helping reconnect a disinterested citizenry to politics. In Wichita, Merritt (1995) reports that voters in the Eagle's circulation area were more aware of campaign issues than citizens elsewhere in Kansas. In Charlotte, according to Miller (1994), Observer readers showed more interest in politics after the newspaper's public journalism initiative began than readers of other area newspapers. And in Madison, Wis., where the local newspaper and public television station staged "town hall" meetings of citizens to discuss campaign issues in 1992 and 1994, about one in three respondents to a telephone survey said the project encouraged them to become more interested in politics (Denton and Thorson 1995: 7).

Nevertheless, a closer look at the evidence yields a less compelling picture. Just one in four readers of the Charlotte Observer noticed anything different about its political coverage after it shifted to a public journalism model (Miller 1994: 67). And a survey taken for the Pew Center for Civic Journalism found that just one in four North Carolina voters had any awareness of a controversial public journalism project built around the 1996 election - an effort that involved regular coverage in 15 newspapers and broadcast outlets throughout the state (Frederick Schneiders Research 1996: 4-5).

During the 1996 election campaign, the Pew Center invited The Bergen Record, a midsized daily newspaper in northern New Jersey, to design a public journalism initiative that included quantitative and qualitative research studying the initiative's effect. Every day except Saturday from Labor Day through the November election, The Record published at least one full page of "public-style" coverage of the contests for president, U.S. Senate, and Congress. With funding from the Pew Center, we conducted focus groups and two waves of statewide surveys that assessed the public impact of The Record's public journalism.

Our work is not intended to judge the ethical arguments advanced in the public journalism debate. Other journalists and scholars make a powerful case that covering politics like a football game fails to engage voters in the political process, and encourages a destructive public cynicism that undermines the foundations of democracy. But our results raise questions about the ability of journalists, acting solely by themselves, to reconnect a disinterested public to politics in a relatively short period of time merely by altering the tone of what they print and put on the air.

There was much at stake for New Jersey voters in the 1996 election. In addition to the race for president at the top of the ticket, the unexpected retirement of Sen. Bill Bradley created the state's first race for U.S. Senate without an incumbent in 14 years. Primary voters in both parties chose veteran members of Congress to run for Bradley's seat. The Democrats nominated Robert G. Torricelli, a seven-term congressman from the heart of The Record's circulation area with a liberal to moderate voting record. Republicans chose Dick Zimmer, a fiscal conservative from the state's western exurbs with moderate views on social issues.

While both men had served in elected office for more than a decade, most New Jerseyans knew little about them; a Star-Ledger/Eagleton Poll taken shortly after the primary showed that only 21 percent of the public had formed either a favorable or unfavorable impression of Torricelli, and just 18 percent had reached an opinion about Zimmer (Star-Ledger/Eagleton Poll 1996). This is frequently the case in statewide contests in New Jersey, thanks to an unusual confluence of political and media structural factors.

On the political side, New Jersey has only three public officials who are elected statewide - the governor and its two U.S. Senators; positions such as attorney general that typically serve as stepping stones elsewhere are appointed by the governor. Moreover, the state has just one commercial television station, and most local television news seen by New Jerseyans originates with New York City and Philadelphia stations that give scant notice to Garden State affairs.

As the state's third-largest daily newspaper, The Record was likely to be an important source of information for many voters. Published in Hackensack, The Record is by far the most widely read daily in Bergen and Passaic counties, a mix of well-to-do suburbs and blue-collar manufacturing towns just west of Manhattan. Its average daily circulation is approximately 151,000 daily and 210,000 Sunday, but market research conducted by Gallup for The Record and other publications in 1995 and 1996 found that over a typical five-day period, some 59 percent of adults in the two counties - that is, about 600,000 people - read The Record at least once (The Record 1996).

The Record was among the first news organizations in the Northeast to introduce public journalism into its political coverage. The newspaper's editor, Glenn Ritt, read about the Charlotte Observer's election project shortly after taking charge of the 300-person newsroom in 1992, and felt New Jersey would benefit from a similar venture. The Record significantly expanded its coverage of issues, beginning with the 1992 presidential contest; widened the focus of The Record Poll to explore public opinion about issues and values in addition to the campaign horse race; and developed regular features, both on the editorial page and in the news columns, that gave greater visibility to the opinions and ideas of readers.

In 1996, the newspaper decided to bring public journalism - which up to that point had been conducted as intermittent special projects - into its pages every day. Beginning on Labor Day, The Record would reserve a full page, without advertising, in the first section of the paper every day except Saturday. Called "Campaign Central," the page would consist entirely of articles about the races for president, U.S. Senate, and Congress which followed a public journalism model.

The page had a standard layout that was designed to look something like a second front page. The "Campaign Central" logo, a stylized star, appeared at the top. It was flanked by teases for pages about the election on The Record's World Wide Web site and for "Instant Feedback," a daily feature of the editorial page that invited readers to phone in their comments about the day's editorials. One or two main stories occupied the center of the page. A strip along the bottom pointed readers to other items in the paper about the campaign and to election-related shows on television. The left-most column contained "Campaign Almanac," a grab bag of news, notes, and quotes from other campaigns around the country.

The "Campaign Central" page for Wednesday, October 16, 1996. The centerpiece story, by Mike Feinsilber of The Associated Press, describes questions that citizens interviewed by AP reporters across the country wanted to ask the presidential candidates. A second story, below, defines programmatic jargon that might have come up during that night's Clinton-Dole debate. Three boxes at the bottom point readers toward opinion articles about the campaign, and preview television coverage. The column on the left of the page, called "Campaign Almanac," contains a variety of political briefs, including schedules for local televised town meetings sponsored by The Record. "Ears" on either side of the main logo refer readers to a chat with New Jersey congressional candidates on the Internet, and promote a feature on the editorial page in which readers commented on the campaign. (Copyright 1996 Bergen Record Corporation. All rights reserved.)

The main articles typically were between 750 and 1,000 words long. About half were written by Record staffers; the remainder came from The Associated Press, the Washington Post, the Los Angeles Times, and the Knight-Ridder News Service. Stories for "Campaign Central" were chosen using four criteria:

• They emphasized issues rather than campaign strategy. Using results from The Record Poll, editors identified the issues that were the most important to New Jersey voters, including job security, crime, education, programs for the elderly, welfare, and the environment. The candidates' positions on each of these issues were explored in detail. Additional articles looked at other issues that were identified by The Record's editorial board, such as the future of the federal judiciary, reform of the Internal Revenue Service, and immigration.

• They gave voice to the concerns of citizens. Mindful of public journalism's call to make citizens the center of political discourse, reporters were assigned to talk to NewJerseyans from different walks of life and explore their views about the challenges facing their state and nation. These stories took several different forms. For example, using a cluster analysis from The Record Poll, one week-long series looked at the values that divide New Jerseyans. Another series took one of the newspaper's local columnists to social clubs, church basements, and school gymnasiums for conversations with groups of local residents.

• They provided more detailed information about the candidates. When journalism is driven by sound bites, voters find it harder to get the information they need to make judgments about candidates, and candidates have more difficulty making their points to voters. Thus many "Campaign Central" stories sought to provide readers a more extensive picture of the candidates, either through extended interviews, verbatim excerpts from speeches and debates, or by giving citizens an opportunity to mail in questions for candidates to answer in their own words. A regular feature called "On The Air" printed the scripts of campaign commercials and sifted the evidence offered by candidates to support their claims.

• They provided opportunities for citizens to participate in the campaign process. In addition to changing the tone of news coverage, many proponents of public journalism also feel that news organizations should reach into their communities and create opportunities for citizens to engage the political process, such as "town hall" meetings with elected officials (Charity 1995; Rosen 1996). Understandably, this has been one of the most controversial points about public journalism, since it strikes some observers as community organizing by a different name (see, for example, Eisner 1994; Conte 1996: 824-825). In conjunction with "Campaign Central," The Record organized a series of public forums about the campaign, open to anyone who returned a coupon published on the page. One session gave citizens an opportunity to discuss the issues they felt the campaigns should address. Other forums featured debates between the congressional candidates, with citizens, rather than a panel of reporters, asking the questions. All the sessions were recorded for telecast on cable, and excerpts from transcripts were published in the newspaper.

The notion of an entire page dedicated practically every day to "public-style" election stories was, so far as we know, without precedent at an American daily newspaper. However, most of the stories themselves - and certainly the criteria the editors used to select them - probably would have seemed familiar to editors of similar projects at other news organizations. The page was intentionally constructed to reflect the prevailing mainstream thinking among practitioners of public journalism - as reflected in the Pew Center's seminars and newsletters - about the elements of a successful election project (see, for example, Thames 1995, Still and Iverson 1995).

"Campaign Central" was intended to be a floor, not a ceiling, for The Record's public journalism commitment; "public-style" stories about the campaign continued to appear elsewhere in the newspaper. By the same token, unlike Charlotte's editors in 1992 (Miller 1994: 37), Record editors continued to print stories in other sections that followed the more conventional, "horse race" style of campaign journalism. They felt readers should have a choice between the two approaches, and that both types of stories should compete on equal terms for space on the front page. Whenever space allowed, strategy and campaign stump stories were grouped together on a page adjacent to "Campaign Central."

But no matter what else was in the paper, the "Campaign Central" concept guaranteed that there would be at least one full page of public journalism about the election published every day. Moreover, "Campaign Central" always received a plug somewhere on the front page - usually a one- or two-sentence abstract of the day's main story, often accompanied by a color photograph.

Much of the material published in "Campaign Central" was brought together in a concise form as an eight-page "Voters' Guide" published the Sunday before the election. This special section was constructed almost entirely in a simple chart format that compared the major candidates side by side. The main chart for each campaign covered biographical information, positions on key issues, voting records, and rankings by special interest groups for the Republican and Democratic candidates. There were brief descriptions of independent candidates, and references to World Wide Web sites with more information about the campaigns. The guide also reprinted a series of Record editorials that reviewed the election's most important issues.

In addition, most "Campaign Central" stories were posted in a special section on The Record's web site. This section was maintained after the election, and can be viewed at http://www.bergen.com/campaign.

The "Campaign Central" project presented a unique opportunity to assess the impact of public journalism. Because New Jersey does not have a single dominant statewide newspaper, it was readily possible to construct a statewide sample of adults who experienced the same campaigns for president and U.S. Senate but saw different daily newspapers. This allowed us to address an issue that challenged other researchers: the absence of a meaningful control group.

Our quantitative design - described in greater detail below - called for two random-digit-dialed telephone surveys of the New Jersey electorate. One wave was completed before the traditional kickoff of the campaign on Labor Day. A second wave was conducted following the election.

We also conducted four focus groups in Bergen and Passaic counties during the seven days immediately before and after the election. These groups, which typically lasted 90 minutes to two hours, enabled us to explore the impact of specific incidents and articles, and provided important anecdotal background to support and explain the quantitative results.

Each group included between 10 and 12 respondents, divided as evenly as possible between men and women. The first two groups were organized by preference in the race for U.S. Senate. Group 1 consisted of regular Record readers who had made up their minds in the Senate race since Labor Day, while Group 2 consisted of Record readers who were still undecided five days before the election. The second two groups were organized by newspaper readership. Group 3 included Record readers who also read at least one other newspaper, while Group 4 consisted of local residents who do not regularly read any newspaper. Participants were paid a modest incentive.

HYPOTHESES

If public journalism works as its advocates propose, what should we expect it to accomplish? At a 1994 conference, Rosen offered a broad but simple test. "In committing an act of public journalism," he said, "you know you have succeeded when you have left behind something people continue to use, some added ability the community now possesses. The power of the press thus empowers others besides the press" (quoted in Charity 1995: 160).

The writings of Rosen and others suggest that this empowering of the electorate will have three dimensions:

1. Better information about politics and policy. If successful, public journalism should promote more effective learning about candidates and issues. Citizens exposed to public-style coverage should be more aware of who the candidates are and be able to describe more completely and accurately the differences between them.

2. Greater public attachment to the political system. By refocusing election coverage on the concerns of regular citizens rather than the gamesmanship of consultants and other insiders, public journalism should strengthen the relationship between politics and people. Citizens who experience public-style coverage should be less cynical about politics and government, and more likely to believe that the political system adequately represents them and their values.

3. Wider public involvement in political decisions. Successful public journalism, supporters say, will do more than change attitudes: it will change behavior as well. Citizens exposed to public-style coverage should be more likely to participate in the decision-making processes of their community. Not only will voting rates rise, but citizens will seek out more opportunities for what Rosen calls "deliberative dialogue" - that is, situations in which people come together to "sort through difficulties, reflect on choices, listen with care, and deepen their views" (Rosen 1996: 50).

Beyond these three dimensions of impact, Rosen also believes that public journalism will have an additional side benefit for the press: it will encourage citizens to hold reporters and their news organizations in higher esteem. Conventional ways of covering politics, he argues, leave citizens so disinterested that they see no reason to read newspapers or watch newscasts. By inspiring citizens to greater involvement in public life, he contends, news organizations that practice public journalism stimulate fresh demand for the information they supply.

In short, as a result of The Record's experiment in public journalism, we would expect the following:

H1. Record readers will be more interested and engaged in the campaign for Senate than New Jerseyans who read other newspapers.

H2. Record readers will be more knowledgeable and demonstrate evidence of a more thorough decision-making process than readers of other New Jersey newspapers.

H3. Record readers should be more participative than readers of other newspapers. They should vote in greater numbers, and have more discussions about politics outside the home.

H4. Record readers should have more positive attitudes about the political process and actors in the process than others. We would expect them to feel that the system is responsive. We would also expect them to have a higher opinion of The Record than of other news organizations.

FOCUS GROUP RESULTS

The focus groups, which began as the final "Campaign Central" pages appeared, left the Record editors who attended them feeling stunned and somewhat shaken. In all three groups containing Record readers, it was clear that most respondents were unfamiliar with the section, even after several sample pages were passed around the room. The frames through which respondents viewed the campaigns seemed to have been shaped mostly by the candidates' commercials. What little news coverage respondents remembered was largely an afterthought.

Each group began with a general discussion of the campaign and the candidates. The conversation invariably brought out widespread feelings of disconnection from and disgust with politics. There was little enthusiasm expressed for any of the choices, and considerable unhappiness with the negative tone, particularly in the race for Senate.

When the discussion turned to important issues dividing the Senate candidates, most answers clearly drew on the campaigns' television commercials. Said "Janet," a 59-year-old secretary who said she reads The Record seven days a week: "The only thing we keep hearing is Megan's Law. That's the only thing I keep seeing that pops up on TV."

Indeed, when asked where they got their information about the campaigns, newspaper readers and nonreaders alike agreed that television - and particularly the commercials aired by the candidates - was most important.

"In the Senate race, I really don't pay much attention, to tell you the truth. I just look over it. I've never, ever talked about it with any of my friends. I watch commercials. That's about it." ("Jason," 22, a graphic artist. Reads The Record five days a week.)

"I'm not learning intelligently. I'm just picking it up from the blitz that's on the TV." ("Laura," 58, a pianist. Reads The Record three days a week.)

Few group participants noticed a difference between The Record's reporting of the 1996 elections and its coverage in previous years. Many described the paper's coverage in words that conveyed little knowledge, depth, or commitment - "fine ... adequate ... tolerable." Like "Donald," a 45-year-old municipal employee who reads the paper five days a week, they often had trouble explaining what they meant.

MODERATOR: What about [The Record's] news coverage of the election...?

"DONALD": At times they were right on the nose, and at times they were out in never-never land.

MODERATOR: Can you give me examples?

"DONALD": Not specifically. There were times you looked at it, and you said, "Hey, that's pretty interesting." And there were times like, who wants to hear this bull-again?

Most participants did not remember seeing a daily page about the election, even once they were reminded of its name. The only session at which more than two or three respondents recalled seeing "Campaign Central" was the group consisting of people who read at least two daily newspapers - hardly a representative sample.

Each group was shown eight examples of articles, charts, and other features about the campaign - a mixture of "public-style" stories and graphics from "Campaign Central" and more conventional coverage from other parts of the paper. The stories were distributed as full-size photocopies of the entire newspaper page, so participants saw them in the context in which they originally appeared.

Once again, few respondents remembered any of the articles, whether public or traditional in style. None of the stories tested produced much interest when the moderator asked whether they were something participants wanted to read.

The charts, however, drew a markedly different response. Participants were much more likely to remember having seen information about the candidates that was presented in chart form, such as a side-by-side grid comparing the environmental positions of Zimmer and Torricelli. They said charts were significantly easier than prose to understand.

"It's laid out line by line, where it's a little more concise, so you can read it better, which is a lot better than some of the articles that - you print the whole thing and then what did it say?" ("Carla," 49, a medical office worker. Reads The Record seven days a week.)

"I like it. It makes it easy to see. You don't have to go reading pages." ("Jon," 37, a car dealer. Reads The Record no more than one day a week.)

Moreover, respondents said that information about the campaign was more believable when presented as a chart than in paragraph form as a news story.

"That chart ... was not a news feature that was written. That was factual." ("Jane," 57, part-time accountant. Reads The Record and/or another newspaper seven days a week.)

"It's no opinions - just reporting. I don't need somebody to tell me what it means. I can read." ("Max," 58, unemployed. Reads The Record two days a week.)

The preference for charts was especially pronounced in the two groups held after the election, in which the "Voter's Guide" came up spontaneously in discussion before the moderator had a chance to mention it. "Jane" referred to the guide as her campaign "bible," a sentiment shared by many other respondents.

"It's like what you get before an exam." ("Victor," 58, a salesman. Reads The Record and/or another newspaper seven days a week.)

"You don't have to go from Page 3 to Page 14 to finish the story." ("Nancy," 39, a hospital worker. Reads The Record and/or another newspaper four days a week.)

Respondents said the "Voter's Guide" charts were so useful that they should have been reprinted continuously throughout the campaign, without concern that information which had already been printed was being repeated.

"I'd like to see them put it out right after the convention ... and then come back to it at the end." ("Max," 58.)

With this exception, however, the focus groups left the distinct impression that "Campaign Central" and its "public-style" election coverage had made hardly any impact on North Jersey voters - especially when compared to the apparently substantial effect of the candidates' TV advertising. But the sample was small, and there was always the possibility that The Record's public journalism had affected citizens in ways they did not or could not remember. For a more comprehensive analysis, we turned to our quantitative data.