Content-Type: text/html PARTISAN AND STRUCTURAL BALANCE OF ELECTION STORIES ON THE 1998 GOVERNOR'S RACE IN MICHIGAN By Frederick Fico and William Cote A paper submitted to the Mass Communication and Society Division of AEJMC for consideration for presentation at the annual convention in Phoenix, Arizona, in August 2000. ABSTRACT The partisan and structural balance of newspaper stories covering the 1998 governor's race in Michigan was assessed and compared to the newspaper coverage of three earlier elections. The 1998 election coverage favored the Democratic challenger in terms of space and prominence given his campaign's assertions. A detailed issue analysis, however, suggests that the Republican incumbent was able to dominate the substantive issue agenda, while the Democratic challenger became himself the issue because of his insulting campaign comments. However, reporters gave more news attention to challengers than to incumbents in each of the four elections compared, regardless of party affiliation. Challengers were favored in six of 10 stories across the four elections, with incumbents dominating fewer than three in 10. The structural balance of stories -- assessed by the degree to which they were constructed to favor one candidate in terms of prominence and total space -- varied with the type of news organization the reporters worked for, with sources reporters used for stories, and with the creative story-construction decisions the reporters made in the writing process. Specifically, stories written by newspaper statehouse bureau reporters were more structurally balanced than those produced by AP or newsroom-based newspaper reporters. Stories based on interviews were more balanced than stories based on activities such as speeches or rallies. Several institutional and newsroom factors indirectly influenced story balance. The bigger the newspaper staff size, the more statehouse bureau stories were produced, which in turn resulted in more balanced election stories. Further, the bigger the newspaper staff size, the bigger the newspaper's total story output, which was related indirectly to more balanced stories. Finally, the bigger the newspaper's staff size the bigger the newspaper's output of stories on the governor's race, which in turn increased the likelihood that stories would be based on interviews that produced more balanced stories. INTRODUCTION News coverage of elections clearly matters powerfully in any society in which citizens use candidate information to guide their voting. But the normative assumption that voters act on the basis of candidate views presupposes that such voters are first exposed to those views. This study explicitly assesses news coverage of those views. Journalists advocate "fairness and balance" in covering public policy issues,1 and news organizations depend economically on a mass audience attracted in part by the credibility of their claims to report fairly.2 Therefore, this study explores how fairness and balance values translated into objectively measurable prominence and space accorded candidate views in election stories. To give context to this research, data from news coverage of four elections are assessed to determine if any news bias related more to the party or to the incumbency status of candidates. Finally, this study analyzes how theoretically related factors such as news organization resources, newsroom priorities, and reporter work norms influenced such election stories. At a practical level, identifying these influences can illuminate what journalists can - or should - change in election reporting to better match their normative values of fairness and balance. CONTEXT OF THE RESEARCH Obviously, ethical norms such as fairness and balance can influence journalists' professional behavior. But such norms may provide little specific guidance for journalists on a particular story. For instance, codes accepted by the American Society of Newspaper Editors, the Society of Professional Journalists and many newspapers called for fairness and balance or related values such as impartiality and unbiasedness.3 But little or no guidance is given on how to impartially or fairly cover a conflict. Similarly, editors have indicated that impartiality is a major component of quality news reporting, and while these journalists may know such quality when they see it, it is unclear what rules they apply to assess particular stories.4 The Concern for Bias Much journalistic concern for fairness and balance in reporting stems from the assumption that lapses in such qualities will lead the public to perceive bias, thereby harming news organization credibility.5 But some journalistic biases are explicitly grounded in professional news values such as impact, proximity, timeliness and "newness" that may result in stories that are arguably unfair because these other values take precedence.6 What journalists define as news, the sources they use and the views from those sources that they present may well constitute "biased samples" when compared to results from more systematic research techniques. For example, one study that included news coverage of Gulf War and abortion issues found that anti-war and pro-choice sources dominated coverage even as public opinion polls indicated more popular support for the pro-war and pro-life positions.7 Certainly reporters possess biases embedded in their individual backgrounds and intellectual processes.8 But such individual biases must still conform to or operate within the institutional constraints and goals of the media organizations themselves.9 Reporters must construct stories that reflect news values. Reporters select sources based on news value considerations of their institutional power, prominence or activity, and on their personal qualities of credibility, accessibility and reliability.10 Further, reporters decide what space and prominence to give those sources in the structure of a story, and in particular what "leads" or characterizes the news value of the story as a whole. And in a conflict, reporters decide whether and how to get the "other side" into a story, and if so, what space or prominence to give them.11 Editors reinforce or inhibit such story construction decisions with their own decisions on use and placement of the resulting stories.12 Election Coverage Bias Many studies of presidential election reporting have generally found even-handed treatment of Republican and Democratic candidates.13 Other studies have illuminated imbalanced news attention to presidential candidates, exemplified by studies by Stoval who found balanced treatment of candidates in 1980 but not in 1984.14 Various explanations for such imbalances have included ownership patterns and liberal bias. Studies focused on incumbency have also found uneven treatment of candidates.15 Clarke and Evans found news bias in congressional elections favoring incumbents, which they attributed in part to the incumbents' greater ability to get news attention. Fico et. al. studying newspaper coverage of statehouse, congressional, U.S. Senate and U.S. presidential races found coverage favoring incumbents at the statehouse and congressional levels, but diminishing at the U.S. Senate and presidential levels. But Lowry and Shidler considered the preponderance of negative sound bites on network TV aimed at Republicans in the 1992 and 1996 presidential elections more consistent with an explanation of liberal bias than of candidate incumbent or challenger status. RESEARCH QUESTIONS AND HYPOTHESES Election Story Characteristics This research assumes that story structure manifests how journalists value fairness and balance. Consequently, this research examined the typical structure of election stories. Were such stories more likely to be one-sided or two-sided in their presentation of candidate views? Within the structure of a story, how prominently and extensively were such views presented? Which candidate dominated particular issues and what did that dominance look like? And more broadly, was coverage more related to the party or to the incumbency status of the candidates? A further assumption in this research is that story structural characteristics may influence the perception of bias by readers. So this study calculated how often stories with typical structures occurred in the election news studied. If readers judge the impartiality of a news organization from just a subset of its campaign stories, what were the chances that they encountered fair and balanced ones in the 1998 governor's election? The following research questions therefore probe the characteristics of these stories: RQ1: How fair and balanced were election stories about the candidates? RQ2: What issues favored candidates, and how did issue stories present candidates? RQ3: How did partisan balance vary by candidate party or incumbency status? RQ4: How structurally balanced were the campaign stories, whatever their partisan tilt? Influences on Election Story Structural Balance Finally, this research explores factors that might influence story balance, and assesses their importance relative to one another. Shoemaker and Reese suggest a hierarchy of influences on news content in general, with higher-level influences constraining factors at lower levels.16 Specifically, they point to societal ideology and other, non-media social institutions as having the broadest influence, followed by media organizational goals and resources, followed by newsroom routines and finally by the individual values and characteristics of media workers. What, then, were the important direct and indirect influences on election story structure? Although this conceptual ordering is useful, little research has been carried out to identify important influences on such a content quality as story balance, or to illuminate the interrelationships of variables at different levels that might influence such a content outcome. However, several influences can be identified on the basis of past research and a priori logic that could be expected to influence content generally, and specifically the election reporting story qualities that are the focus of this research. At the organization level, the staff available to cover an election campaign can certainly be expected to influence content outcomes. Certainly bigger staffs can be expected to produce more stories about an election. Further, several studies have found that management's allocation of staff resources to specialized bureaus can affect the quality as well as quantity of a news organization's public affairs reporting. Legislative studies by Fico found that reporters assigned to bureaus were more likely to use more sources in stories than wire service reporters or reporters working from local newsrooms.17 However, more staff resources, all else equal, should reduce the pressures on individual reporters to fill available news hole. Consequently, the following hypotheses are suggested: H1: News organizations with more staff will produce more election stories by specialized bureaus than will news organizations with fewer staff members. H2: News organizations with more staff will produce more election stories than will news organizations with fewer staff members. H3: News organizations with more staff will produce fewer election stories per reporter than will news organizations with fewer staff members. Previous research on statehouse bureau reporting suggests that such specialized reporters use sources differently than do their newsroom-based colleagues or wire service reporters. Specialists obviously develop more of a knowledge of sources, and may have developed relationships that make those sources more accessible when need for stories. If such reporting qualities are also applicable to the statewide governor's race stories produced by such bureau reporters, the following hypotheses are plausible: H4: Election stories produced by statehouse bureaus will be more likely to rely on interviews than stories produced by newsroom reporters. H5: Election stories produced by statehouse bureaus will be more likely to be two-sided than stories produced by newsroom reporters. H6: Election stories produced by statehouse bureaus will be less likely to be imbalanced than stories produced by newsroom reporters. Media sociology studies indicate news work is shaped by newsroom influences that editors directly control. Editors communicate work norms for story numbers and type of coverage.18 It is plausible that reporters who must produce more stories in a limited amount of time may compromise by using fewer sources in each story. One possible result is fewer stories relying on interviews, which require more initiative and time than easily observable rallies or other campaign events. Similarly, hard-pressed reporters may produce more one-sided and imbalanced election stories compared to colleagues who are able to take more time on fewer stories. Therefore, the following hypotheses relating editorially imposed constraints to election reporting are suggested: H7: The more election stories reporters produce, the less they will be based on interviews. H8: The more election stories reporters produce, the less they will they include both sides. H9: The more election stories reporters produce, the more imbalanced they will be. METHOD This study explores these questions and hypotheses with a content analysis of all hard-news stories on the 1998 governor's race in Michigan's nine largest dailies from Labor Day to election day.19 The study analyzes campaign stories containing quoted or paraphrased assertions by Democratic and Republican candidates and their supporters.20 The partisan balance of the 1998 governor's race stories was compared to similarly analyzed ones produced by the same news organizations covering the 1994 Michigan governor's race and the 1996 U.S. presidential and senate races.21 The 1998 election pitted incumbent Republican Gov. John Engler against Democratic nominee Geoffrey Fieger. Engler, running for his third term as governor, ultimately won with about 63 percent of the vote. Fieger, most widely known previously as defense attorney for assisted suicide advocate Jack Kevorkian, had won an upset victory in a three-way race for the Democratic nomination. Fieger was dogged during the campaign for past and more recent comments that were construed as anti-Christian and anti-Semitic. He also was criticized for campaign comments ridiculing the governor's family and personal appearance. Fairness and Balance Measures Story fairness is defined as the presence of quoted or paraphrased assertions by sources supporting both Republican incumbent Engler and Democratic challenger Fieger. The proportions of one-sided and two-sided stories therefore indicate the probability of reader exposure to stories they might plausibly construe as unfair or fair. The partisan balance of each story was assessed by first determining the partisan balance of each of four components. These are: (1) total space: did the two candidates obtain equal story space for their assertions, or did one get more?; (2) lead position: did both candidates get to make assertions in the story's first paragraph lead, or was one candidate included but not the other?; (3) paragraphs 2-5 position: did both candidates get to make assertions somewhere in the story's second through fifth paragraphs, or was one candidate included but not the other?; and (4) paragraphs 6-10 position: did both candidates get to make assertions somewhere in the story's sixth through tenth paragraphs, or was one candidate included but not the other? For example, the lead position component would be judged as (1) favoring Engler, if Engler-support assertions were included but not Fieger's ; (2) favoring Fieger, if the reverse were the case; (3) balanced, if both candidates' assertions were included in the lead; or (4) irrelevant, if neither candidates' assertions were in the lead. The overall partisan balance was then determined by whether more of the relevant components favored one candidate than the other. A related index of each story's structural balance uses these same components to assess the degree to which a story was dominated by the same candidate. It was created by adding the number of balance components dominated by Engler, subtracting those dominated by Fieger, and then taking the absolute value of that figure. A 0 therefore indicates a structurally balanced story (and also one balanced between Engler and Fieger on the partisan index) and a 4 indicates that the same candidate dominated every structural balance component. Validity and Reliability of Balance Measures The validity of these balance measures depends on the assumption that stories are read from the top down and that material higher up has a greater, but unknown, likelihood of being encountered than material lower down.22 The measurement weighting given assertion prominence therefore flows from the logical probabilities of a reader's exposure to story material. An arguably balanced story that deals sequentially with candidate assertions may not be perceived as such by a reader who stops before the paragraph in which the opposition begins to speak. Quantifying partisan assertion position and length in the structure of stories avoids validity problems from categorizing the partisanship of whole stories or relying on evaluations of "tone." A story classified as favoring the Republican or Democrat, for instance, may arguably still be fair in that it included assertions from the opposition. And while judgments of news bias may be reliable in the context of a research study, such judgments are of unknown validity when it comes to what average readers may perceive. Focusing instead on the measurable relative treatment of opponent assertions does not require assumptions about the way readers may judge them. A two-person coder reliability assessment for this study was performed on approximately 5 percent of stories randomly sampled from all relevant ones. All variables with less than 100 percent agreement still achieved a Scott's Pi of at least .90. The alpha reliability for the components of the partisan and structural balance indices was .78. Explanatory Variables Election stories were also coded for variables that could help explain variation in story structural balance drawn from three of the hierarchical levels suggested by Shoemaker and Reese. At the organizational level, each newspaper's staff size was recorded as a measure of resources that could affect election coverage. Further, each newspaper's stories were also coded for their statehouse bureau or local newsroom origin as a measure of different institutional norms governing reporters even within the same news organization. At the newsroom level, each governor's race story was coded for that newsroom's total general story and specific election story production as a measure of overall work norms and election coverage priority. As a control tapping editorial priority for election stories, each governor's race story was coded for front page, section front page or inside placement. At the individual reporter level, similar counts were made for the total election story production of each reporter as measures of that reporter's work norm. The reporter's use of interview sources in election stories was also considered a measure of independence in covering the campaign. The reporter's total production of all stories was included as a control. Analysis of Data These data constitute the universe of what was available and findings therefore do not require inferential statistics for generalization. Research questions on partisan balance of stories are addressed with proportions. Questions on structural balance of stories are addressed with mean balance scores. Regression analysis was used to assess hypothesized influences of explanatory variables on the balance of the stories. Significance tests were used in this analysis only to identify more substantively important influences on story structural balance. RESULTS Some 400 stories were relevant for analysis from the nine newspapers. About 47 percent of these stories were produced by Associated Press, a third by local newsroom staff members and the balance by statehouse bureau reporters. About 15 percent of stories were run on a section front page and about 17 percent were run on Page One. About 56 percent of the stories were based on interviews. About a third of all the stories were one-sided. Partisan Balance of Stories Overall, Fieger dominated the partisan balance of 51 percent of stories on the 1998 governor's race compared to 35 percent for Engler. He got more space in 51 percent of the stories compared to 33 percent for Engler. Fieger dominated 21 percent of leads compared to Engler's 18 percent. Fieger dominated paragraphs two through five in 34 percent of stories compared to 24 percent for Engler. Fieger-support assertions were therefore more lengthy and more prominent than those supporting Engler, and thus more likely to be encountered by readers. Fieger dominated coverage in every newspaper, but by widely varying margins. He dominated the Flint Journal by just 2 percentage points and the Saginaw News by just 4 percentage points. By contrast, Fieger led in the Oakland Press by 32 percentage points and in the Grand Rapids Press by 39 percentage points. Campaign Issues Coverage of particular campaign issues also varied from the overall pattern in the sample. Nine substantive issues were coded during the 1998 governor's race. Engler dominated three issues -- the economy, corruption, and education - that totaled 15 percent of the sample. Fieger dominated four issues - law enforcement, the budget, candidate personality and "other" -- that made up 53 percent of sample stories. Two issues, the environment and welfare, were evenly balanced between the two candidates, but made up only 7 percent of sample stories. In addition, Fieger dominated a tenth, "horse race" issue, which made up about a quarter of the sample. Stories on the most covered issues dominated by each candidate, making up nearly 40 percent of the total sample, display significant differences in overall story partisanship and the extent to which candidates dominated story agendas. Education, covered by 39 stories, was the largest one dominated by Engler. Ten randomly selected education stories examined in detail were highly partisan. On average, two-thirds of the stories' paragraphs were dominated by partisan assertions. And 80 percent of the leads contained such assertions, twice the overall sample average, indicating that the candidates mostly set the story agendas on this issue. Seven stories from five of the nine sample newspapers covered Engler's proposal for "freedom schools" in which parents would be given more local control. The extensive attention given this proposal, even though Engler himself admitted that he had few specifics, probably resulted from the governor's previous legislative success in enacting sweeping education finance reform. Fieger forces charged the proposal would make the job of educators even more difficult. Some stories also discussed whether Engler's earlier reform of education finance had more helped or hurt school districts. Other stories concerned school dropouts and teacher strikes. Most of the 10 stories used non-partisan as well as partisan sources, either to provide data or to interpret the issue. Candidate personality, covered by 113 stories, was the largest one dominated by Fieger. Some 28 randomly-selected personality stories were on average less dominated by partisan assertions than the education ones. Overall, only about half of all story paragraphs contained partisan assertions, and only 57 percent of story leads contained them. But this masks a significant division in the topics of such personality stories. A substantial minority of stories examined in detail the candidates' backgrounds and values. For example, the lead of an Oct. 5 story in the Saginaw News asserted: "Whatever you think of Geoffrey Fieger as a gubernatorial candidate, there's no denying he's a dynamite attorney." In addition, obviously pre-planned personality stories on candidates were paired, often side-by-side on the same page, dealing with such the candidates' religious faiths and the perspectives of their spouses. Most stories, however, focused on Fieger's vitriolic election campaign assertions. For example, he called the governor a "bully and coward " for opposing assisted suicide in an Oct. 1 story in the Kalamazoo Gazette, and during rallies Fieger was commonly cited referring to the size of the governor's "gluteus maximus." In other stories, Fieger had to defend himself against criticism for such assertions. For example, an AP story in the Grand Rapids Press on Oct. 28 reported Fieger explaining as "comedic" his comment that Engler was the product of a human- animal mating. In the same story Fieger denied ever saying that the governor's children might have tails. A sidebar detailed Fieger quotes about the governor, call-in show listeners' questions about them, and Fieger's responses. In addition, stories focused on the Engler-campaign's attacks on Fieger's character and fitness for office. If these analyzed stories are typical of the of the campaign coverage, then Engler was able to set the issue agenda of the election even when vehemently contested by opponents who got more say. Fieger's personal attacks on the governor often made himself the issue. And stories in which Fieger had to defend his attacks were easily as numerous as stories giving readers information and insights into his beliefs and values. Fieger certainly got more news attention than the governor, but at least some of that attention may have been wasted or even self-defeating. Party and Incumbency The novelty of Democratic challenger and the outrageousness of some of his comments may alone have been sufficient to obtain more attention than the incumbent Republican governor during the 1998 campaign. But stories written on the 1994 governor's race and on the 1996 U.S. Senate and presidential races indicate a similar pattern of attention to challengers. Specifically, stories written by newspaper staff and statehouse bureau reporters on the four elections suggest that the incumbency or challenger status of the candidates was more influential than party in determining coverage balance. The four-election average indicates that Republican and Democratic candidates each dominated about four in 10 stories. (See Table 1) But in each election, the challenger was heavily favored in the coverage, dominating six in 10 stories on the four elections compared to incumbent domination of three in 10 stories. (See Table 2). Interestingly, Fieger obtained less coverage for his assertions than did any other challenger. Structural Balance of Stories The structural balance of these stories assesses more generally than partisan balance how much candidates dominated stories, whatever their party or incumbency status, and therefore may indicate how likely readers are to perceive fairness or bias in the news. In general, the data show that a reader of a story dominated by either Engler or Fieger was not very likely to encounter prominent or extensive assertions by the opponent. Each candidate dominated stories by more than two of the four balance components measured. Overall, readers had nearly identical 4 in 10 chances of encountering stories that were either mostly balanced (0-1 on the scale) or mostly imbalanced (3-4 on the scale). Story fairness and structural balance were obviously related. About 35 percent of stories were both fair (i.e., two-sided) and mostly balanced, while 21 percent of stories were both unfair (one-sided) and mostly imbalanced . Influences on Story Structural Balance Hypothesized relationships for staff size all were supported. Specifically, staff size was strongly related to a newspaper's total production of governor's race stories (Beta = .72), to the production of statehouse bureau election stories (Beta = .27), and to the production of election stories by individual reporters (Beta - .26). However, statehouse bureau story production was not significantly associated with interview sourcing or with the use of both sides in stories. Statehouse bureau stories were associated, however, with less imbalance in election stories (Beta = -.20). Finally, hypothesized relationships for reporter election story production were not supported. Contrary to expectations, the more stories a reporter wrote on the election, the more likely they were to be two-sided (Beta = .17). Table 3 displays the influence of all variables at various levels, including the control variables, on the dependent variable of story balance. Figure 1 illustrates the causal interrelationships among variables hypothesized to have influence on story balance. The most important direct influences on story structural balance, controlling for the effects of all the other variables, were at the organization and reporter levels. (See Table 3) However, the creative story construction decisions of reporters accounted for the most important direct influences on story structure balance. Stories that included both campaign sides were much more likely to be balanced than one-sided stories. And stories constructed with a lead partisan assertion were much more likely to be imbalanced than stories without such a prominent partisan assertion. These relationships should not be surprising. A story that does not include one of the sides cannot possibly be balanced. And given that a story's lead sets the "agenda" for the whole story, it is not surprising that a lead partisan assertion predicts more such assertions in the story paragraphs directly following. The model of these relationships illuminates the need to take into account how higher-level factors indirectly influenced lower-level ones that in turn affected story structure. (See Figure 1). Clearly the larger the newspaper staff, the more stories they produced generally, the more stories they produced specifically on the governor's race, and the more those governor's race stories were produced by statehouse bureaus. These in turn directly or indirectly diminished imbalance in the resulting election stories. Newspapers producing a greater volume of governor's race stories than others were also more likely to have reporters who based those stories on interviews. Given a news organization's commitment to election coverage, it seems plausible that editors would support broader source searching attempts by reporters, resulting in relatively more interview stories that in turn diminished story imbalance. DISCUSSION AND IMPLICATIONS News coverage of the 1998 governor's race in Michigan could be perceived by readers as significantly imbalanced. Readers had only a 1 in 3 chance of encountering stories that were fair and balanced, as assessed by measures in this study. In fact, a third of stories were completely one-sided, and 4 in 10 stories were mostly imbalanced, whether they were fair or not. Voters trying to make up their minds about candidates did not encounter very many individual stories giving even-handed presentations of their differences on issues. But it may be impossible for news coverage of any election to be "perfectly" fair and balanced in terms of the measures used in this study. And given journalistic news value priorities that may depend on some particular circumstance, that outcome may not even be desirable. In this election, Democratic challenger Geoffrey Fieger made personal attacks that were both newsworthy themselves and that in turn aroused newsworthy critical reaction. The governor, by contrast, avoided personal confrontation and even stayed away from any debates with the challenger. The result was substantial news attention that focused on Fieger, attention that the detailed personality issue assessment suggests may not have served his campaign well. In fact, in the context of all four elections observed, Fieger got less attention than any other challenger. Still, these findings lend more support to the notion that news bias follows candidate institutional status rather than journalistic partisanship. Further, the direction of that bias indicated in this research was to favor challengers. Possibly challengers had this advantage because these elections were highly visible and statewide in scope, with challengers taking more energetic measures to get attention and news media more willing to give "unknowns" more exposure. An incumbency advantage may still hold at smaller-constituency electoral levels. Organizational and newsroom priorities had important effects on story structure. Organizations large enough to deploy reporters to statehouse bureaus obtained more balanced election stories as a result. And in general, the more reporters a newspaper had, the more they could be used for election coverage which in turn may have given reporters enough experience in election coverage to be more sensitive to story balance. The newsroom priorities of editors also made important differences in the election coverage. News organizations concerned with fairness and balance might deploy more staff to election reporting and provide support for more reporter initiative in election story sourcing. Specifically, when editors gave reporters more time and encouragement to do interview stories, the result was more fair and balanced stories. Moreover, the lead emphasis of those interview stories was more likely to be neutral or balanced than stories based on activities. The reporters' creative control over the construction of stories was a major influence on the structural balance assessed in this research. Even fair stories could be badly imbalanced in the structure used for presenting information. Journalists who wish to be fair and balanced may therefore be undermining their own efforts in the story construction process if one candidate is given unchallenged prominence in placement and attention. And more broadly, enhancing the manifest fairness and balance of election stories could help improve the credibility of the news media as a source of unbiased public policy information crucial to democratic decision-making. ENDNOTES 1. See Louis A. Day, Ethics in Mass Communication. (Belmont, Calif.: Wadsworth Pubg. Co., 1991), 38-39. Several of these codes are also listed in the appendices of William L. Rivers, Wilbur Schramm and Clifford Christians, Responsibility in Mass Communication (N.Y.: Harper & Row, 1980). 2. See Cecilie Gaziano, "How Credible is the Credibility Crisis?" Journalism Quarterly 65 (Summer 1988): 267-278; 375; Cecilie Gaziano and Kristin McGrath, "Measuring the Concept of Credibility," Journalism Quarterly 63 (Autumn 1986): 451-462. 3. "Be fair, unbiased, accurate, complete, factual, professional, aggressive, and compassionate," the ASNE advised in "Newspaper Credibility: Building Reader Trust." (Washington, D.C.: American Society of Newspaper Editors, April, 1985), 11. The Freedom Forum announced Feb. 24, 1997 that it would allocate $1 million to "explore new procedures to enhance free press, to increase fairness in the media and to improve public perception of fairness in the media." See "Freedom Forum News," Vol. V, No. 3, Feb. 24, 1997, 1. 4. Accuracy and impartiality of reporting ranked as first and second attributes of editorial quality in a 1977 survey of 746 editors cited in Leo Bogart, Press and the Public. 2nd ed. (Hillsdale, N.J.: Erlbaum, 1989). 5. See: Christine D. Urban, "Why Newspaper Credibility Has Been Dropping: A report for the American Society of Newspaper Editors." Urban & Associates, Inc. December 1998. 6. Most reporting texts emphasize both the values of conflict and fairness. See Brian S. Brooks, George Kennedy, Daryl R. Moen and Don Ranly, News Reporting & Writing 4th ed. (New York: St. Martin's Press, 1992) 3-15. 7. Frederick Fico and Stan Soffin, "Fairness and Balance of Selected Newspaper Coverage of Controversial National, State and Local Issues." Journalism and Mass Communication Quarterly 72 (Autumn 1995): 621-633. 8. S. Holly Stocking and Paget H. Gross, How Do Journalists Think? (Smith Research Center, ERIC Clearinghouse, 1989). 9. Pamela J. Shoemaker and Stephen D. Reese, Mediating the Message: Theories of Influences on Mass Media Content. 2nd ed. (White Plains, N.Y.: Longman, 1996). 10. Angela Powers and Frederick Fico, "Influences on Reporters' Use of Sources at High Circulation U.S. Newspapers." Newspaper Research Journal 15 (Fall 1994): 87-97. 11. Tuchman used the term "strategic ritual" to characterize the journalistic convention of getting both sides of a controversy into a story in order to avoid criticism. See: Gaye Tuchman, "Objectivity as Strategic Ritual," American Journal of Sociology 77 (January 1972): 660-679. 12. Sociologist Warren Breed referred to the "newsroom socialization" of reporters as they learned to conform to conventions rewarded or punished by superiors. See: Warren Breed, "Social Control in the Newsroom," Social Forces 33 (May 1955): 326-335. 13. Guido H. Stempel III, perhaps the most tenacious investigator of how prestige newspapers cover presidential elections over the years, has found roughly equal attention given to Republican and Democratic candidates. For his most recent studies, see: Guido H. Stempel III and John W. Windhouser, "The Prestige Press Revisited: Coverage of the 1980 Presidential Campaign," Journalism Quarterly, (Spring 1984); 49-55; Guido H. Stempel III and John W. Windhauser, "Coverage of the 1988 Presidential Campaign." Journalism Quarterly 70 (Summer 1993): 311-320. 14. See examples of such mixed results in: James Glen Stovall, "The Third Party Challenge of 1980: News Coverage of the Presidential Campaign," Journalism Quarterly 62 (Summer 1985): 266-271; James Glen Stovall, "Coverage of the 1984 Presidential Campaign," Journalism Quarterly 65 (Summer 1988): 443-449; 484. Bias attributed to ownership ideology was found by Keith Kenny and Chris Simpson, "Was Coverage of the 1988 Presidential Race by Washington's Two Major Dailies Biased?" Journalism Quarterly 70 ( Summer 1993): 443-449; 484. Johnson found a front-runner bias in coverage of the 1988 Democratic primary. See: Thomas J. Johnson, "The Seven Dwarfs and Other Tales: How the Networks and Select Newspapers Covered the 1988 Democratic Primaries." Journalism Quarterly 70 (Summer 1993): 311-320. Mantler and Whiteman found a pro-Republican bias in newspaper coverage of the 1992 presidential race, while Lowry and Shidler found an anti-Republican bias in network news coverage of that race. See Gordon Mantler and David Whiteman, "Attention to Candidates and Issues in Newspaper Coverage of the 1992 Presidential Campaign." Newspaper Research Journal 16 (Summer 1995): 14-28; Dennis T. Lowry and Jon A Shidler, "The Sound Bites, the Biters and the Bitten: An Analysis of Network TV News Bias in Campaign '92." Journalism and Mass Communication Quarterly 72 (Spring 1995): 33-44. 15. Peter Clarke and Susan Evans, Covering Campaigns. (Palo Alto, Calif.: Stanford University Press, 1983); Frederick Fico, John Clogston and Gary Pizante, "Influences of Candidate Party and Incumbency on News Coverage of the 1984 Elections in Michigan Dailies." Journalism Quarterly 65 (Autumn 1988): 709-713; Dennis T. Lowry and Jon A. Shidler, "The Sound Bites, the Bitters, and the Bitten: A Two Campaign Test of the Anti-Incumbent Bias Hypothesis in Network TV News," Journalism and Mass Communication Quarterly 75 (Winter 1998): 719-729. 16. Shoemaker and Reese, op. cit. 17. Frederick Fico, A Comparison of Legislative Sources in Newspaper and Wire Service Stories. Newspaper Research Journal 5 (Spring 1984): 35-44; Frederick Fico, The Replication of Findings on newspaper and Wire Service Source Use in Two Statehouses. Newspaper Research Journal 7 (Winter 1985): 45-51. 18. Frederick Fico, Perceived Roles and Editorial Concerns Influence Reporters in Two Statehouses. Journalism Quarterly 62 (Winter 1985): 784-790; Frederick Fico, Influence of Perceived Editorial Concern and Role Self-Concept on Source Use of Reporters in Two Statehouses. Journalism Quarterly 63 (Summer 1986): 322-330. 19. Newspaper circulations range from 50,000 to more than 300,000 and account for about 70 percent of Michigan's total daily circulation. They are the Saginaw News, the Macomb Daily, the Kalamazoo Gazette, the Lansing State Journal, the Oakland Press, the Flint Journal, the Grand Rapids Press, the Detroit News and the Detroit Free Press. 20. Several types of stories were excluded such as reporter analyses, and "Q & A" stories. Poll stories or horse race assessment were excluded unless they also contained assertions by candidates or their supporters. Stories about Gov. Engler in his role as governor also were excluded unless campaign assertions were present or unless such stories indicated that some governmental action had campaign implications. Assertions were quoted or paraphrased statements explicitly linked to candidates or their supporters by verbs of attribution indicating speaking such as "said," "stated," "argued," etc. Verbs denoting state of minds of candidates or their supporters (e.g., "feels," "thinks," "believes," etc.) were also considered to signal assertions when it was clear that such verbs were used as synonyms for verbs such as "said." 21. See: Frederick Fico and William Cote, "Fairness and Balance in Election Reporting," Newspaper Research Journal 18 (Summer/Fall, 1997): 50-63; Frederick Fico and William Cote, "Fairness and Balance of Stories in Newspaper Coverage of the 1996 Presidential Election." Journalism and Mass Communication Quarterly. 76 (Spring 1999): 124-137. Data on the coverage of the U.S. Senate race has not been reported elsewhere. The 1994 governor's race pitted Gov. Engler against Democratic challenger Howard Wolpe, a former congressman. The 1996 U.S. Senate race saw Republican challenger Ronna Romney defeated by Democratic incumbent Carl Levin, while in the presidential race Republican Bob Dole, the former U.S. Senate majority leader, was defeated by President Bill Clinton. 22. Many studies have estimated readership based on article position or size in a news medium, but none have estimated declines in readership as readers move down through a story. Fiedler et. al. in their reporting text estimate that more than 40 percent of readers fail to go beyond the fifth paragraph of a story, but no research documentation is given for this estimate. See: Fred Fedler, John R. Bender, Lucinda Davenport and Paul E. Kostyu, Reporting for the Media. 6th ed. Fort Worth, Texas: Harcourt, 1997) :107. Table 1: Balance of Stories Favoring Republicans and Democrats (Percent of Newspaper-Produced Stories* Favoring Candidates) Gov. Sen. Pres. Gov. All 1994 1996 1996 1998 Favors Republicans 28% 63% 61% 33% 43% Favors Democrats 58% 17% 30% 53% 44% Balanced 14% 20% 10% 14% 13% Story N 214 76 221 210 721 Table 2: Balance of Stories Favoring Challengers and Incumbents (Percent of Newspaper- Produced Stories* Favoring Candidates) Gov. Sen. Pres. Gov. All 1994 1996 1996 1998 Favors Challengers 58% 63% 61% 53% 61% Favors Incumbents 28% 17% 30% 33% 26% Balanced 14% 20% 10% 14% 13% Story N 214 76 221 210 721 * Stories exclude AP stories in each election year, and in 1996 also exclude syndicated stories from other sources covering the presidential race. Table 3: Direct Influences on Story Structural Balance (Coefficients are Betas. Statistically significant Betas are in bold. A positive relationship indicates that the influence increases imbalance. A negative relationship indicates that the influence decreases imbalance.). Influence Type Story Balance Organization Level Staff Size .05 Statehouse Bureau -.20 Newsroom Level Story Prominence -.01 Total Story Production -.06 Total Governor's Race Stories .00 Reporter Level Reporter's Total Stories -.06 Reporter's Governor's Race Stories .02 Interview-Based Story -.18 Story-Construction Level Two-sided story -.46 Partisan Lead Story .44 Explained Variance = .57 Equation Significance = .00 Figure 1: Model of Direct and Indirect Influences on Story Structural Balance. (Coefficients are statistically significant Betas. A positive sign indicates the influence increases story imbalance. A negative sign indicates the influence decreases story imbalance.) Statehouse Bureau Two-Sided Election Story Story Newspaper Staff Reporter's Total Story Size Election Stories Imbalance Newspaper's Total Interview Election Stories Story