Content-Type: text/html Inability to Recognize News Source Bias and Perceptions of Media Bias Donna Rouner Michael D. Slater Judith M. Buddenbaum Colorado State University Department of Technical Journalism Fort Collins, Colorado 80523 970-482-2967 Email: [log in to unmask] June 15, 1995 To be presented at the annual meeting of the Association for Education in Journalism and Mass Communication, Newspaper Division, Washington, D.C., 1995. Professors Rouner, Slater and Buddenbaum are associate professors in the Department of Technical Journalism Colorado State University. We wish to thank the students in the methodology and research classes who assisted on this project. Inability to Recognize News Source Bias Inability to Recognize News Source Bias and Perceptions of Media Bias Abstract Some scholars and journalists would argue American journalism in the 1990s suffers from a credibility crisis. This paper examines public perceptions of the news media and the bias of news media sources, comparing newspaper employees and their audience. Using three samples, a general public sample of adults (N=245), a college student sample (N=172), and a newspaper employee sample from a local newspaper (N=28), this study found key differences between the journalism professionals and their audience relat ive to perceptions about news media providing balanced news and using quotes from sources to show two sides of an issue. Further, the audience differences from the professionals in the ability to perceive source bias predicted audience members' views that news media are less likely to be balanced and use two-sided quotes from sources. Inability to Recognize News Source Bias Inability to Recognize News Source Bias and Perceptions of Media Bias Some scholars and journalists would argue American journalism in the 1990s suffers from a credibility crisis (cf., Merritt, 1995). Public perceptions of bias in the news media were detected as early as the 1960s (Roper, 1969). However, trend data suggest the American public views the media as increasingly less trustworthy, paralleling--possibly surpassing--a downward slide in trust for other American institutions (Harris, 1984; American Society of Newspaper Editors, 1985). What does it mean, specifically, when public opinion polls show an increasing distrust for newspapers, television and other news media? Do members of the public perceive the news media as out of touch, dishonest, biased? How do public perceptions correspond to journalists' perceptions? Is it part of the downward trend in trust in the media, along with other American institutions, a function of younger people's general cynicism and mistrust? This paper explores these questions. We examine perceptions of the role of the news media and perceptions about bias, specifically in the use of news sources by public affairs journalists. Perceptions of the general public, college students, and of media personnel are explored and compared. More controversially, we propose that there is, and provide evidence for, a link between the naivete of news consumers with respect to source bias and their belief that media are biased in their news coverage. Inability to Recognize News Source Bias 2 Bias and the News Media Although bias is viewed as an important dimension of source credibility (Hovland and Weiss, 1951, McCroskey, 1966), it is understudied. Credibility research about news media more often explores the importance of expertise as an attribute of news sources or a newspaper, or comparatively between newspaper and electronic news sources (Gaziano and McGrath, 1986; Meyer, 1988; Rouner, Brown and Day, 1993). However, Hovland and Weiss (1951) conceptualized expertise and trustworthiness as the two dominant dimensions of source credibility at the inception of social science inquiry into this topic. Even trustworthiness, however, is conceptually problematic with respect to judgments about the news media. One might expect trust in the news media to be a function to some extent of the degree to which the media are perceived to be objective or biased. Therefore, this study is concerned with perceptions of bias on the part of news media, news consumers' perceptions of bias on the part of news sources, and their possible relationship. Some government officials and political pundits have criticized what they perceive to be a systematic ideological bias in the media, particularly the news media. Empirical evidence suggests that the "liberal bias" charge against the media is largely unfounded (c.f., Entman, 1989). Some scholars argue, however, that at certain times the American press has been more liberal or conservative, reflecting the climate of the country at the time (Bennett, 1988; Rosen, 1992; Taylor, 1992). Psychological research has found that media bias is sometimes a mote in the eye of Inability to Recognize News Source Bias 3 the beholder. In particular, researchers have found that partisan or biased individuals from both sides of a controversial issue are likely to perceive the same media coverage as biased in opposite directions (Vallone, Ross and Lepper, 1985). Gunther and Lasorsa (1986) found that greater importance individuals placed on issues led to their increasing trust of newspaper coverage on that issue. Gunther (1988), however, found a curvilinear relationship between attitude extremity and trust in media coverage of an issue, suggesting some consistency with the Vallone et al. (1985) findings. It would be, of course, self-serving on the part of journalists to claim that news consumer partisanship is the major reason for accusations of media bias. Bennett (1988), for example, provides a typology of four types of bias in the news that causes journalism harm in the eyes of the citizenry. First, he argues that media personalize news into human interest accounts, limiting the public's ability to see the "big picture," and causing a focus on trivial aspects of important news events, like personality flaws and behavioral gaffes. Second, he argues that a drama bias causes news media to present stories that stem from events, leaving no professional convention for addressing many of the most serious problems confronting contemporary societies, like hunger, racism, resource waste and depletion. A third type of bias is information fragmentation, again making it difficult to see larger issues. Finally, Bennett describes a source bias, where news media seek out authoritative voices of officials who offer views that normalize the news for members of an average public. Bennett's claim regarding source selection is supported by the work of a number of Inability to Recognize News Source Bias 4 researchers. Most of the research on news and public affairs information bias centers on the use of selected, elite sources (Sigal, 1973; Gandy, 1982; Stempel and Culbertson, 1984; Brown, Bybee, Wearden and Straughan, 1987). Sigal (1973) suggested that reporters rely on sources who are easy to identify and access, with most front-page news in the New York Times and the Washington Post based on government sources. Brown et al. (1987) examined how elites dominate as news sources, relied on because they are readily available and thought to provide trustworthy, authoritative information in an articulate fashion. They found, similar to Sigal (1973) and others, that government sources--primarily males in executive positions--dominated. Perceptions of News Source Bias and its Possible Relation to Perceptions of Media Bias Accusations of systematic bias are frustrating and painful for journalists. Reporters are trained to operate under a professional code of values and ethics, where fairness and balance reign supreme (Charnley and Charnley, 1979). All the while, audiences are showing increasing signs of distrust. Some of that distrust may be due to extreme partisan positions on the part of some consumers. This, however, is hardly likely to explain a more widespread questioning of media objectivity. The structural limit ations of journalism, and the patterns of sources typically used, may constrain and in some senses bias news coverage. These problems and limitations in contemporary journalism may dispose news consumers to be critical of the news media. News consumers, however, have also changed; newspaper readership is shrinking and as a consequence there may be less attentive and less expert consumers of news. It may be that there is something in the way journalists approach Inability to Recognize News Source Bias 5 achieving objectivity that news consumers do not acknowledge, misunderstand, or reject. The problem of journalistic objectivity, of course, is a particularly difficult one. The Oxford American dictionary defines objectivity as "not [being] influenced by personal feelings or opinions" (Oxford, 1980). While journalists may seek to exclude their own feelings or opinions from a story, they are dependent upon sources--who usually have such opinions--in developing the story. News stories, then, are inevitably influenced by feelings and opinions, even when not the journalists' own. Journalists, in most reporting contexts, finesse this problem through the convention of story balance. This practice may often mean purposefully seeking out biased sources to present one side's view and the other side's view. This is, of course, presented in the name of balance and fairness; it is not objective, but rather presents two biased sources' perspectives. The end goal is, presumably, neutrality. Some journalists argue this might be construed as an unusual way to present unbiased information (Merritt, 1995). Journalists may actually be relying on the most influential, active sources they can find, given the various constraints of reporting, rather than purposefully looking for a biased perspective (Ericson, Baranek and Chan, 1987). Clearly, there is more to the use of sources from two sides of an issue in the name of fairness and balance that leads to perceptions of bias or distrust in the news. However, as Lacy, Fico and Simon (1991) point out, fairness or balance in any given news story is typically a matter of story balance. They measured balance by determining if contact was made with someone representing the two major sides of a controversy. In their study of news stories in 21 newspapers, they found many lacked balance. In 28 percent of instances Inability to Recognize News Source Bias 6 the other side of a controversy was not contacted, and no explanation was offered by the story's writer. In an additional 6 percent, no contact was made, but an explanation suggested efforts for such contact failed. They measured balance by counting the number of words presented by each side and creating an absolute difference score. Nonetheless, the extent that the news media are to be perceived as unbiased must depend in part on perceptions of story balance, as well as the actual frequency of adequate balance in news coverage. This study concentrates on the extent and consequences of discrepancies between journalism professional's and the public's perceptions of news source bias. We define bias as a perceived attribute of a news source whereby the individual, or the group the news source represents, has a clear, vested interest in a cause or action relative to changing or maintaining the status quo. This bias is apparent knowing the person's credentials. Those credentials are presented such that one might predict how that individual would respond ideologically on political and social issues. Thus, if background information describes an individual's positions and accomplishments but that information does not in and of itself suggest a position on social or political concerns, the source is interpreted as low in bias. A highly biased source would be generally presented in a journalistic account as operating to alter the status quo in some fashion, like eliminating taxes. Source bias as a concept is continuous. Any source, no matter how it is presented, would convey some degree of bias, as in an allegiance to some group or cause, or a tendency toward action. However, for purposes of this study, a dichotomy of high and low will be used. The Lacy et al. (1991) results suggest one obvious problem, a discrepancy between Inability to Recognize News Source Bias 7 the journalistic value of presenting two sides of an issue and the actual practice of doing so. However, in this study we propose that there may also be a more fundamental problem. The success of the convention of story balance assumes that the news consumer shares the reporter's assumptions concerning the bias, or lack thereof, of the various sources cited in the study. This assumption has received little empirical examination. One recent study suggested that, at least among college students, there was very little ability to recognize and acknowledge bias among message sources (Slater & Rouner, 1993). It may well be that many members of the general public are relatively naive with respect to judging the extent of bias of news sources. On the other hand, journalists have extensive opportunity in working with news sources to experience first-hand the extent of their bias or self-interest. We might expect, then, that: Hypothesis 1: There will be differences in perceptions of news source bias between professionals in the news industry and a) members of the general public and b) young adults. In particular, we might expect to find that professionals in the journalism industry will more readily attribute bias to sources with a strong probable self-interest related to the news story. The nature of the differences with respect to sources with relatively less self-interest is more problematic. On one hand, members of the public may be simply less attuned to bias judgments, and more inclined in the case of both high and low bias sources to make neutral judgments. On the other hand, members of the public may tend to assume less bias than the more experienced and perhaps more cynical journalism professionals even among relatively less apparently self-interested sources. Inability to Recognize News Source Bias 8 What might be the consequences of such a mismatch between journalist and public perceptions of bias? The problem is most easily apparent in the case of high bias sources. Story balance is achieved by offering the views of competing biased sources in counterpoint. If the bias of these sources is not well-distinguished by some news consumers, they are unlikely to grasp the nuances of story balance. Such news consumers, then, should be more likely to perceive bias in the news media. Therefore: Hypothesis 2: The less bias that a) members of the general public and b) young adults perceive in high biased sources, relative to journalism professionals' own judgments, the greater their perceptions of media bias. This, of course, represents a counterintuitive prediction. In the absence of the process described above, one would expect that members of the public who tend to perceive bias in media institutions would also be quick to attribute bias to individual news sources as well. After all, such sources are also normally affiliated with social institutions, and ones with clear political and social agendas of their own. The issues around perceptions of low-bias sources are more complex. On one hand, members of the public may be, as discussed earlier, more likely to make neutral judgments about sources. On the other, they may tend to consistently perceive less bias among low bias sources as well as among high bias sources. Either case, though, might under different circumstances reduce the ability to recognize and correctly process the story balance convention. Sometimes, a low-bias source is used to provide presumably objective comment on an issue. If news consumers perceive the source as biased, then the intended effect of increased objectivity is lost. If less-biased sources are used to provide opposing viewpoints, Inability to Recognize News Source Bias 9 the inability to perceive potential bias inherent in embracing a given viewpoint will interfere with acceptance of the story balance convention. Therefore: Hypothesis 3: Differences between journalism industry professionals and a) the general public and b) college students in perceptions of bias among apparently less-biased news sources will be associated with greater perceptions of media bias. Of course, all these hypotheses are based on the assumption that members of the public are more critical of the news media and have different expectations regarding the news media than do journalism professionals. These expectations and assessments are also compared in this study. Method Data Collection Three samples were generated for this study, a convenience sample of N=172 students enrolled in an undergraduate journalism course at a Western University, a random sample of N=245 individuals over 18 who live in a small Western community near the university, and a mail sample of N=28 employees of the newspaper serving that community. In November, 1993, the students responded to a call for volunteers and completed a survey instrument en masse in their classroom setting. Uninterested students engaged in an alternative activity. This yielded a 95% response rate. Trained telephone interviewers asked nearly identical questions to the general audience sample in March, 1994, with a 67% response rate. Inability to Recognize News Source Bias 10 The same survey instrument that was used for the student sample, with a few minor changes in demographic questions, was dropped in the office mailboxes, or left with clerks for distribution, in September, 1994, at the newspaper under study. Employees used stamped return envelopes to mail the instruments back to the researchers. Two follow-up letters served as reminders for the employees to complete the surveys and were left, with additional survey instruments, in the employee mailboxes. Twenty-eight of the 72 employees responded, a 39% response rate. Measurement All individuals responded to a series of questions about information they get from the news media in general, from newspapers, television, radio and news magazines. First, they were asked to rank their perceived roles of journalism, with a "1" as the most important role and "5" as the least important role, using each number only once. The roles included: to entertain people, to interpret events, to make a profit, to provide neutral, unbiased news coverage, and to serve as a watchdog. They were also asked to respond to statements regarding their level of agreement, using a Likert-type 5-point response scale that ranged from strongly agree to strongly disagree. These statements were, "I can trust the information I get from news stories," "Most news stories present issues accurately," "News stories are just that--stories. They are not factual accounts of events," "Most news stories are presented in a balanced and fair manner," "News stories usually quote at least two sources--one from e ach side of an issue," "Most sources quoted in news stories are knowledgeable," "Sources quoted in news stories Inability to Recognize News Source Bias 11 are usually appropriate to the story," and "Most sources quoted in news stories are just trying to bring attention to themselves or to the cause they represent." To investigate specific differences in perceptions about the use of biased sources, eight descriptions were generated as source stimuli. Similar to experimental conditions that manipulate bias in the source credibility literature, four types of sources were described in individual paragraphs. Political, scientific, religious and business-related sources were described as either low or high in bias, depending on information provided about their allegiances, affiliations and background experiences. Names that were used (or gender references in the telephone sample of the general public) were male, to avoid a sex bias confound. Information about the individual's affiliations and experience was included in order to convey a degree of bias for each individual. The four high bias descriptions follow: A spokesperson for Capital and Class--a citizen lobby that works to change the way federal and state governments in the U.S. operate. Founded in 1970, Capital and Class is committed to developing a Marxist tradition within movements of liberation in the U.S., including the labor, anti-racist, environment, peace and women's movement. A researcher in environmental ethics at the University of California at Berkeley, who also heads the California chapter of Earth First. He has been arrested seven times for chaining himself to trees to prevent logging in old growth forests in the Pacific Northwest. He recently received a $70,000 grant from the National Endowment for the Arts to build an art museum display commemorating the northern spotted owl. A graduate of the University of Connecticut and former $500,000-a-year investment banking executive at JBMI Financial Corporation, who was convicted of banking and securities fraud in 1990. Although he is now in prison, he devotes his time to writing on a wide range of social and economic topics. His latest book, "Taxpayers, Suckers and Free Riders," was published by the Libertarian Press. Inability to Recognize News Source Bias 12 A clergyman who is the head of mail order ministries at the Church of "What's Happening Now." He holds college degrees in marketing, mass communications and religious studies and recently wrote a book titled "Churches as Tax Dodges." His television show can be seen on Sunday mornings in Kansas City, Missouri. The four low bias descriptions follow: A man who was a tenured professor in international political science at Michigan State University and who is now a private consultant to U.S. and foreign government trade organizations. His areas of expertise include world population growth and environmental economics. The chairman of the Department of Immunology at Penn State who regularly contributes to the Journal of the American Medical Association. He has served with distinction on the boards of several different scientific organizations, including the Smithsonian Institution's Science Information Exchange. A professor of comparative religion at Arizona State University. He is the author of more than 30 books, including "Religious Social Doctrine and New Thinking in Economics," and "That They May Speak: Voices of Religious Social Thought." He also contributed articles to The Encyclopedia of World Religions and Religion in America. The chief financial officer at Brown Investments International, a publicly traded company operating in nine states and two foreign countries. He holds graduate degrees in economics and medical ethics from the University of Illinois. He has written more than 20 articles on economics and social issues and is a regular contributor to The Journal of the American Economics Association. The general audience sample responded to a question that asked how biased they considered each of the eight possible sources to be, using a 7-point scale that ranged from unbiased to biased. The students and newspaper employees reacted to six of these source stimuli, presented in random order on the instruments, by responding to a question that asked how biased they considered each of the possible sources to be. The remaining two sources were each presented in relationship to a text that the respondents reacted to. Responses included Inability to Recognize News Source Bias 13 with a text are not analyzed in this study to keep bias measures consistent. Additional questions ascertained media use patterns and demographics for the three samples. The media use questions, about which media are the most important sources for news and information, differed slightly among the three groups. For this exploratory study, simple frequency percentages and mean comparisons were used to assess patterns of similarity and differences among the three samples. Because the professional sample is small, even the independent sample T-tests used to determine significant differences should be interpreted with caution. We are primarily concerned with patterns of difference, given the exploratory nature of this project. The more important function of the professional sample was to provide a benchmark with respect to bias assessments of the sources against which we could compare our general public and college student samples. In order to create an index for each respondent that represented a discrepancy between public perception of bias and journalist perceptions of bias, difference scores were calculated for the perception of each source stimuli in the general public and student samples. For the high biased source stimuli, each respondent's raw score measuring the perceived bias of a source stimulus was subtracted from the mean of the journalists' assessment of the same source stimulus. This was done for the eight source stimuli. For the low biased source stimuli, the mean of the journalists' assessment of a source stimulus was subtracted from each respondent's raw score measuring perceived bias of the same source stimulus. Again, this was done for all eight source stimuli. The difference in the directions corresponds to the Inability to Recognize News Source Bias 14 predicted differences in perceptions. A higher discrepancy score on a high biased source meant less bias perceived by the general public sample compared to the newspaper employee sample. For low biased source discrepancy scores, the higher the score, the more the distance between the general public and newspaper employee sample, and between the students and the newspaper employee sample. These discrepancy scores were added to form indices of high source bias discrepancy and low source bias discrepancy for the general public and the students (Cronbach alpha coefficients: General public high source bias discrepancy alpha= .65; general public low source bias discrepancy alpha=.66; students high source bias discrepancy alpha=.62, students low source bias discrepancy alpha=.50[1]. Results To generally describe the samples, the students were 45% male and 55% female, closely corresponding to the university population breakdown. About one-third said they were journalism majors, with 30% majoring in another liberal arts area and the remainder in hard science, agriculture, forestry, etc. Television was the preferred medium for the news (62%), with magazines the least preferred (54%). The mean education level for the general public sample was some college, the mean age 54. Forty-five percent were male and 55 percent female, reflecting the conventional female bias of a telephone sample. The primary news source of the general public sample was listed as television for international news (73%); television for national news (74%); television for state news (34%), followed closely by their local newspaper (29%); and their Inability to Recognize News Source Bias 15 local paper for local news (87%). The newspaper employee sample (of the newspaper referred to above as the local paper) was 67% female, with an mean age of 38. The mean longevity at that newspaper was about 5 years; 34% were in news editorial positions. Twenty percent were in advertising, while 15% were in production. The mean educational level was a college degree. The medium used as a primary source for news was newspapers (64%) followed by television (29%). The medium least used as a source for news was magazines (50%). In assessing the role of journalism in American society, the three samples showed strong similarity on their most important role choice. The most respondents (65% of the general public, 69% of the students, and 86% of the newspaper staff) selected as the most important role providing neutral, unbiased news coverage. The only other role for which more than 10% of any of the respondents from their respective samples selected as first in importance was to interpret events. None of the newspaper employees, however, selected the interpretation of events as the most important (i.e., the number one in importance) role of the mass media. As for the student sample, 11% indicated this was the number one most important role, and 9% of the general public selected this as most important. The student sample emphasized this role across the second and third choice, compared to the others, with 83% indicating it was either first, second or third in the five roles described. This compares to 68% of the newspaper employees and 56% of the general public who placed it in the top three positions of roles. Inability to Recognize News Source Bias 16 Regarding the role of the news media's serving as a watchdog, none of the newspaper employees indicated this role as most important (ranked 1 out of 5). However, 50% of them placed it as second in importance. This compared to 18% of the general public indicating this was the most important (ranked 1 out of 5) and 31% selecting this as second in importance (ranked 2 out of 5), with a total of 49% for the top two categories--close to the newspaper employees' 50%. A little less than 5% of the students chose the watchdog role as first in importance (ranked 1 out of 5), but 30% selected it as second in importance (ranked 2 out of 5), with only 35% reporting this in the top two positions. The three groups were quite similar in how they ranked providing entertainment as a role of news in the mass media, which was about in the middle (3 out of 5 ranked positions). Students were quite different in their views about the news media making a profit. While 32% of the general public and 25% of the newspaper employees placed making a profit at first, second or third as the most important role for news media, only 13% of the students ranked profiting that high. Conversely, 67% of the students ranked making as profit as last, or fifth, in importance as a role. This compared to 49% of the general public sample and 46% of the newspaper employees. General Perceptions About News Media and News Media Sources Turning to the respondents' general perceptions about the news media, two differences were found between the general public and student perceptions (See Table 1). Inability to Recognize News Source Bias 17 Table 1 Mean Differences Between General Public Sample and Student Sample on Statements about the News and News Sources Newspaper General Sample Student Sample N=245 N=172 standard standard Source Type mean deviation mean deviation T-score Trust news info 2.94 1.07 3.20 .83 ns News not factual 2.53 .60 2.19 .62 3.88** Issues accurate 3.07 1.00 3.24 .85 1.77* News balanced 3.09 1.03 3.10 .88 ns Sources 2-sided 2.86 1.11 2.89 .95 ns Sources informed 3.34 .91 3.40 .78 ns Sources appropro 3.74 .62 3.75 .71 ns Sources biased 3.06 .95 3.02 .87 ns *=p<.05 ** = p< .01 *** = p< .001 The general public was more inclined to perceive the news as just stories, not factual accounts (general public mean=2.53, student mean=2.19, t=3.88, p<.01), and the students were more likely to see issues presented accurately (student mean=3.24, general public mean=3.07, t=1.77p<.05). Otherwise, these two groups showed similar perceptions. When comparing the general public with the newspaper employee sample, more differences resulted (See Table 2). Inability to Recognize News Source Bias 18 Table 2 Mean Differences Between General Public Sample and Newspaper Employee Sample on Statements about the News and News Sources Newspaper General Sample Employee Sample N=245 N=28 standard standard Source Type mean deviation mean deviation T-score Trust News Info 2.94 1.07 3.43 .92 ns News not factual 2.53 .60 1.93 .92 3.21** Issues accurate 3.07 1.00 3.24 .85 1.77* News balanced 3.09 1.03 3.77 .83 3.69** Sources 2-sided 2.86 1.11 3.43 .88 2.90** Sources informed 3.34 .91 3.46 .92 ns Sources appropro 3.74 .62 3.82 .62 ns Sources biased 3.06 .95 3.11 .95 ns *=p<.05 ** = p< .01 *** = p< .001 The same pattern emerged on the items about the news presenting stories, not facts (general sample mean=2.53, employee sample mean=1.93, t=3.21, p< .01) and issues being presented accurately (general sample mean=3.07, employee mean=3.24, t=1.77, p< .05). The newspaper employees perceived the news media as more likely to present facts, not stories, and to present issues accurately. These two groups also showed differences on the perceptions of how balanced and fair the news is and whether sources are quoted from two sides of an issue, with the Inability to Recognize News Source Bias 19 newspaper employees more likely to agree with these assertions (for news balanced, general public mean=3.09, employee mean=3.77, t=3.69, p< .01; for sources are two-sided, general public mean=2.86, employee mean=3.43, t=2.90, p< .01). In a similar fashion, the student sample differed with the newspaper employee sample on these two items, news balance and fairness and whether news stories included quotes from two sides of an issue (see Table 3) (for news balanced, student mean=3.10, employee mean=3.77, t=3.70, p< .001, for sources are two-sided, student mean=2.89, employee mean=3.43,, t=2.81, p< .001). Table 3 Mean Differences Between Student Sample and Newspaper Employee Sample on Statements about the News and News Sources Newspaper Student Sample Employee Sample N=172 N=28 standard standard Source Type mean deviation mean deviation T-score Trust news info 3.20 .83 3.43 .92 ns News not factual 2.19 .62 1.93 .92 ns Issues accurate 3.24 .85 3.24 .85 ns News balanced 3.10 .88 3.77 .83 3.70*** Sources 2-sided 2.89 .95 3.43 .88 2.81** Sources informed 3.40 .78 3.46 .92 ns Sources appropro 3.75 .71 3.82 .62 ns Sources biased 3.02 .87 3.11 .95 ns *=p<.05 ** = p< .01 *** = p< .001 Inability to Recognize News Source Bias 20 Perceptions of Source Bias Turning to the reactions to the source stimuli, indices of the low and high sources were not created for comparison across the three samples, given the low number of newspaper employees. Significant mean differences between the low and high sources within the sample groups provided evidence of concept validity for these manipulated items. Tables 4 , 5 and 6 show the comparative means and t-tests between paired sets of the three samples. The general public and student samples showed no differences (see Table 4). However, differences were found between the general public and the newspaper employees and between the students and the newspaper employees. Table 4 Mean Differences Between General Public Sample and Student Sample on Perceived Bias of Hypothetical Sources General Sample Student Sample N=245 N=172 standard standard Source Type mean deviation mean deviation T-score High Political 5.08 1.60 5.08 1.83 0.00 High Scientific 5.18 1.92 5.30 1.61 .58 High Religious 5.01 1.86 5.09 1.73 .38 High Business 5.18 1.61 5.33 1.64 .79 Low Political 3.81 1.57 3.61 1.63 1.08 Low Scientific 3.44 1.67 3.31 1.63 .65 Low Religious 3.94 1.72 4.07 1.70 .64 Low Business 3.63 1.55 3.80 1.61 .89 *=p<.05 ** = p< .01 *** = p< .001 Inability to Recognize News Source Bias 21 Looking at how the general public and the newspaper employee samples differed on perceived bias of the hypothetical sources (see Table 5), three of the sources manipulated to appear highly biased were perceived differently by these two groups (political source: general sample mean=5.08, newspaper employee mean= 5.80, t=1.87, p< .05; scientific source: general sample mean=5.18, newspaper employee mean=6.33, t=5.53, p< .001; religious source: general sample mean=5.01, newspaper employee mean=6.22, t=4.54, p< .001). The high biased business source was not significant, but its means differed in the hypothesized direction. As predicted in Hypothesis 1, the newspaper employees perceived the highly biased sources to be more biased than the general public perceived them to be. Table 5 Mean Differences Between General Public Sample and Newspaper Employee Sample on Perceived Bias of Hypothetical Sources Newspaper General Sample Employee Sample N=272 N=28 standard standard Source Type mean deviation mean deviation T-score High Political 5.08 1.60 5.80 1.64 1.87* High Scientific 5.18 1.92 6.33 .69 5.53*** High Religious 5.01 1.86 6.22 .95 4.54*** High Business 5.18 1.61 5.67 1.27 1.71 Low Political 3.81 1.57 3.10 1.67 1.83* Low Scientific 3.44 1.67 2.62 1.16 2.89* Low Religious 3.94 1.72 3.68 1.89 .53 Low Business 3.63 1.55 3.74 1.74 .09 *=p<.05 ** = p< .01 *** = p< .001 Inability to Recognize News Source Bias 22 As for low biased sources, the political and scientific sources showed differences (political source: general sample mean=3.81, newspaper employee mean=3.10, t=1.83, p< .05; scientific source: general sample mean=3.44, newspaper employee mean=2.62, t=2.89, p< .05). Here, the newspaper employees perceived the low biased sources to be significantly less biased than the general public. As for comparisons between the student sample and the newspaper employees (see Table 6), the same pattern of findings occurred. Each of the high-biased sources except the business source were significantly different (political source: student sample mean=5.08, newspaper employee mean=5.80, t=1.78, p< .05; scientific source: student mean=5.30, newspaper employee mean=6.33, t=4.73, p< .001; religious source: student mean=5.09, newspaper employee mean=6.22, t=4.15, p< .001) . Although only the scientific low-biased source showed a difference between the two groups (student mean=3.31, newspaper employee mean=2.62, t=2.33, p< .05), all the mean differences in sources, low and high, were in the expected direction. Again, the newspaper employees showed a higher perception of bias than the student sample on the high-biased sources and a lower perception of bias on the lower-biased sources. Inability to Recognize News Source Bias 23 Table 6 Mean Differences Between Student Sample and Newspaper Employee Sample on Perceived Bias of Hypothetical Sources Newspaper Student Sample Employee Sample N=172 N=28 standard standard Source Type mean deviation mean deviation T-score High Political 5.08 1.83 5.80 1.64 1.78* High Scientific 5.30 1.61 6.33 .69 4.73*** High Religious 5.09 1.73 6.22 .95 4.15*** High Business 5.33 1.64 5.67 1.27 1.14 Low Political 3.61 1.63 3.10 1.67 1.30 Low Scientific 3.31 1.63 2.62 1.16 2.33* Low Religious 4.07 1.70 3.68 1.89 .79 Low Business 3.80 1.61 3.74 1.74 .15 *=p<.05 ** = p< .01 *** = p< .001 Discrepancy Between Source Bias Perceptions and its Relationship to Media Credibility The two major differences found between the general public and the newspaper employee samples, and between the student and newspaper employee samples, were the general public and students perceived news media to be significantly less likely than the newspaper employees did to present balanced, fair news and to quote sources from two sides of an issue. Inability to Recognize News Source Bias 24 Given these findings, Hypothesis 3 was tested by regressing the indices of the perception of high source bias on these two newspaper credibility measures, balanced news and quoting two sides. Greater distance between general public and newspaper employee, and between student and newspaper employee, perceptions of source bias are represented by a positive beta coefficient in the regression analysis. In the general sample, the more distance between the perceptions of a high biased source, the more respon dents perceived the news media as presenting quotes from two sides of an issue, with a beta coefficient of .29, p< .001 (see Table 7)[2] . Eight percent of the variance was accounted for (p< .001). The pattern of the remaining coefficients is in the direction of the hypothesis. These data lend support for Hypothesis 2, that the less bias members of the public perceive in high biased sources, relative to the judgments of journalism professionals' own judgments (the greater the discrepancy between members of the public and the newspaper employees), the greater their perceptions of media bias. The same dependent variable, only for low bias source discrepancy, was regressed on presenting balanced news and quoting two sides of an issue. Results supported Hypothesis 4 (see Table 8). For the general public sample, the more discrepant the perceptions of low source bias, compared to the newspaper employees, the less likely the public perceived the news media to Inability to Recognize News Source Bias 25 Table 7 High Bias Source Predictors of Perceptions About News Media Credibility Entries are standardized beta coefficients and significance level Perception About News Discrepancy between General Discrepancy between Students Media Credibility Sample and News Employees and News Employees (N=156) N=170) beta R2 beta R2 Present balanced, fair news .14 .02 .11 .01 Present quotes from two-sides of an issue .29*** .08*** .07 .07 *=p<.05 ** = p< .01 *** = p< .001 Table 8 Low Bias Source Predictors of Perceptions About News Media Credibility Entries are standardized beta coefficients and significance level Perception About News Discrepancy between General Discrepancy between Students Media Credibility Sample and News Employees and News Employees (N=156) (N=170) beta R2 beta R2 Present balanced, fair news -.22** .05** -.16** .16* Present quotes from two-sides of an issue -.23** .05** -.10 .10 *=p<.05 ** = p< .01 *** = p< .001 Inability to Recognize News Source Bias 26 present balanced, fair news (beta= -.22, p< .01) and to present quotes from two sides of an issue (beta= -.23, p< .01). Thus, the greater the difference between the two groups, the greater the perceptions of media bias. For the students, this same prediction held up relative to discrepancy of low bias sources' predicting the perception that news media present balanced, fair news (beta= -.16, p< .01). The nonsignificant finding was in the hypothesized direction. Discussion These results provide support for the hypothesis that members of the public, as well as college students, are less likely to discriminate between low and high bias sources. The results also generally support the hypotheses that suggest that such perceptual mismatching is associated with perceptions of news media bias. Of course, the presence of this association cannot provide evidence that the relationship is causal. However, the fact that respondents who assessed biased sources to be relatively less biased also assessed the news media to be relatively more biased would otherwise be counterintuitive. While we by no means wish to suggest that the possible failure of the story balance convention with some consumers is primarily responsible for perceptions of media bias, these data provide at least some preliminary and provocative evidence that it may play a contributing role. The tendency of respondents from the general public, and among college students, to fail to discriminate news source bias in the same way journalism professionals did has even broader implications for the working journalist. If a sizable percentage of news consumers are not making the same assumptions regarding the bias of self-interested news sources as is the journalist, the Inability to Recognize News Source Bias 27 news story will not be interpreted as it was intended. What must the reporter or the editor do to clarify the biases of the news sources, and the implications of these biases for the story? It is important that future research clarify the extent of the problem with news story interpretation, rather than merely assuming such problems exist because of limited discrimination of news source bias. These results, preliminary though they may be, are troubling. Unsurprisingly, the overall patterns of these results suggest that journalism employees perceive journalism more favorably than their audiences and more favorably than students, including journalism majors. Consistent with the discrepancies in perception of source bias, the general public and students did not perceive balance and the presentation of quotes showing two sides of issues as prevalent in the news media. The finding about the role of journalism in interpreting events, with the students assigning far more importance to this role than the journalism employees or the general sample, suggests another possible problem between what the news media do and what the young readership and viewership wants or believes is important as a media role. The general public sample was particularly low on this, compared to the other two. While the general public may be complacent about the news media's remaining neutral and fair, staying out of news interpretation as much as possible, the younger news audience may see this as what is necessary and missing. Where else does one obtain interpretation? As news media may be less inclined to engage in investigative reporting, interpretation, lengthy accounts, the younger cohort of potential subscribers may see this role as especially important. The results of this study are clearly preliminary. Results are obtained from a single Inability to Recognize News Source Bias 28 location, the sample of journalism professionals is small and is not confined to people working in an editorial capacity, and the influence of news source bias perceptions not directly linked to problems in interpreting balance of actual stories. Were the samples larger, the student sample could be broken down into journalism and non-journalism majors to see if majoring in journalism helps explain the results. However, the results are just as clearly intriguing and provocative. If audience perceptions of news source bias are in fact often inaccurate from the journalist's perspective, the success of the story balance convention may be less certain and news consumer interpretation of stories in some cases more distorted than journalists, and journalism researchers, assume. Inability to Recognize News Source Bias 29 References American Society of Newspaper Editors. (1985). Newspaper credibility: Building reader trust. Washington, D.C.: American Society of Newspaper Editors. Bennett, W. Lance. (1988). News: The politics of illusion. New York: Longman. Brown, J. D., Bybee, C.R., Wearden, S.T., and Straughan, D.M. (1987). Invisible power: Newspaper news sources and the limits of diversity. Journalism Quarterly, 64:1, 45-54. Charnley, M. V., and Charnley, B. (1979). Reporting (4th edition). New York: Holt, Rinehart and Winston. Entman, R.M. (1989). Democracy without citizens. New York: Oxford University Press. Ericson, R.V., Baranek, P.M., and Chan, J. B. L. (1987). Visualizing Deviance: A Study of News Organization. Toronto: University of Toronto Press. Gandy, O. H., Jr. (1982). Beyond agenda-setting: Information subsidies and public policy. Norwood, N.J.: Ablex Publishing Company. Gaziano, C. and McGrath, K. (1986). Measuring the concept of credibility. Journalim Quarterly, 63, 451-462. Gunther, A. (1988). Attitude extremity and trust in media. Journalism Quarterly, 65, 279-287. Gunther, A., and Lasorsa, N. (1986). Issue importance and perceptions of a hostile media. Journalism Quarterly, 63:844-848. Harris, L. and Associates, Inc. (1984). Harris Public Opinion Poll. Hovland, C. and Weiss, W. (1951). The influence of source credibility on communication effectiveness. Public Opinion Quarterly, 15, 635-650. McCroskey, J. C. (1966). Scales for the measurement of ethos. Speech Monographs , 33, 65-72. Lacy, S., Fico, F., and Simon, T.F. (1991). Fairness and balance in the prestige press. Journalism Quarterly, 68, 363-370. Merritt, D. (1995). Public journalism. Presented at the Society of Professional Journalism Annual Regional Meeting, Denver. Meyer, P. Defining and measuring credibility of newspapers: Developing an index. Journalism Quarterly, 1988. Inability to Recognize News Source Bias 30 Oxford (1980). Oxford American dictionary. New York: Oxford University Press. Roper, B., (1969). A ten-year view of public attitudes toward television and other mass media, 1959-1968. New York: Television Information Office. Rosen, J. (1992). Politics, vision, and the press: Toward a public agenda for journalism. In Rosen, J. and Taylor, P. The new news v. the old news: The press and politics in the 1990s. The Twentieth Century Fund, Inc., 1-33. Rouner, D., Brown, R., and Day, A.A. (1993). A Meta-analysis of the concept source credibility. Paper presented at the annual meeting of the Midwest Association for Public Opinion Research, Chicago. Sigal, L. V. (1973). Reporters and officials: The organization and politics of newsmaking. Lexington, Mass.: D.C. Heath and Co. Slater, M.D. and Rouner, D. (1993). Spokesperson expertise and bias: An experiment. Presented at the annual meeting of the International Communication Association, Washington, D.C. Stempel III., G.H., and Culbertson, H.M. (1984). The Prominence and dominance of news sources in newspaper medical coverage. Journalism Quarterly, 61:671-676. Taylor, P. Political Coverage in the 1990s: Teaching the Old News New Tricks. In Rosen, J. and Taylor, P. The New News V. the Old News: The Press and Politics in the 1990s. The Twentieth Century Fund, Inc., 1992, 35-69. Vallone, R.P., Ross, L. and Lepper, M.R. The Hostile Media Phenomenon: Biased Perception and Perceptions of Media Bias in Coverage of the Beirut Massacre. Journal of Personality and Social Psychology, 49:3, 577-585, 1985. Inability to Recognize News Source Bias 31 Endnotes [1] These Cronbach's alpha reliability coefficients are reasonable, given the method in which these indices were created. The original measures were dispersed farther in the process of creating the difference scores. Als o, these scores each represent an index of four sources that are similar in bias but differ according to topic discussed. [2] Sample sizes dif fer here, and in Table 8, because of the complexity of creating the discrepanc y indices. For the general public sample, missing values on any o ne variable led to a smaller sample size. Because the eight high and low bias sources were rotated in the students and the newspaper employee s amples, not all respondents in each sample reacted to the same six source stimuli, and the two with text. Therefore, the student sample regression analyses used a means substitution for missing values command, thought appropriate in this exploratory study.