Editorial Commentary © 2014 AEJMC Reprints and Permissions: Sagepub.Com/Journalspermissions.Nav DOI: 10.1177/1077699014549937 Jmcq.Sagepub.Com

JMQXXX10.1177/1077699014549937Journalism & Mass Communication QuarterlyAuthor Name 549937research-article2014

Editorial

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 645–650 Editorial Commentary © 2014 AEJMC Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014549937 jmcq.sagepub.com

Contributors to Journalism & Mass Communication Quarterly sometimes receive a “desk reject” note from the editor indicating that a submission is outside the purview of the journal. In addition, authors are sometimes given advice before the manuscript goes out for peer review on how to improve the presentation, how to report essential information that reviewers will want to know, and so on. “Making these changes does not guarantee acceptance,” authors are told, “but I have been a contributor, reviewer, associate editor, and now editor of Quarterly for thirty-five years, and I have a pretty good sense of what our reviewers expect in a manuscript.” * * * Thirty-five years ago, as an assistant professor trying to juggle my career while starting a family and helping my wonderful wife make a home, I spent a lot of eve- nings at home on my research. It was the pre-Internet, pre-digital reference index era. If you wanted to track down an author or an article, you went to the library stacks and looked at paper indices. One year, perhaps for my birthday, my wife DeeDee gave me a set of cumulative indices to Journalism Quarterly, as this journal was known until 1984. I don’t know where she got them—perhaps from AEJMC (Association for Education in Journalism and Mass Communication) headquarters. The first, published in December 1948, covered volumes 1 to 25 and included a brief history of the Quarterly based on notes from the journal’s four editors (Lawrence W. Murphy, Frank Luther Mott, Ralph D. Casey, and Raymond B. Nixon), along with notes on the forerunners of the journal. The second cumulative index was published in summer 1964, just before Nixon retired after twenty years as editor, and covered volumes 1 to 40 (1924-1963). Guido H. Stempel III was editor when the next two cumulative indices, for volumes 41 to 50 (1964-1973) and volumes 51 to 60 (1974-1983), were published. When the next one was published in 2000 (for the years 1984-1993, or volumes 61-70), I was an associate editor of Quarterly, and Jean Folkerts was editor. Paper indices are outdated, most would argue. Mine are faded, and the binding glue is cracked on two of them. But that set of Quarterly indices permitted me, on one occasion, to leaf through the pages and discover the names of scholars and thinkers who had published in the journal, some of them people I might not even have imagined looking for with one of today’s data base search engines. Most Quarterly readers would be unsurprised by many names on my “discovery list”; they realize already that top contemporary and recent scholars in our field published in Quarterly—including the late Jim Carey and Steve Chaffee, Mel DeFleur, Max McCombs, the late Ev Rogers, Pam Shoemaker, Phil Tichenor, the late Jim Tankard, and Dolf Zillmann.

Downloaded from jmq.sagepub.com by guest on December 20, 2014 646 Journalism & Mass Communication Quarterly 91(4)

But the indices also let me discover historical data on who had published in Quarterly that I would not have ever discovered otherwise. Try some of these names: Mortimer J. Adler, Bernard Berelson, Edward Bernays, Willard G. “Daddy” Bleyer, Warren Breed, Paul Deutschmann, Rudolf Flesch, George Gallup, George Gerbner, Morris Janowitz, Elihu Katz, Harold Lasswell, Paul Lazarsfeld, Nathan Maccoby, Ralph McGill, H. L. Mencken, Frank Luther Mott, Ralph Nafziger, Elisabeth Noelle-Neumann, Elmo Roper, Wilbur Schramm, Eric Sevareid, Percy Tannenbaum, and David Manning White. Quarterly has been home to thinkers and scholars from a broad sweep of disciplines and approaches, including a philosopher like Adler, a satirist like Mencken, pollsters like Gallup and Roper, or a Pulitzer-winning editorialist like McGill. The breadth of content and approach in Quarterly and the journal’s purpose—that is, to serve readers whose interests ranged from public relations to law to history to effects to ethics to economics, to name a few—forced contributors to write and editors to accept work that had appeal to the broad readership. As students and scholars, we read Chaffee and Carey and Janowitz and Rogers and Schramm and Tichenor and Tannenbaum, not because of some narrow topical interest, but because what they wrote had broad theoretical application, value, and appeal. More scholars in our field should aspire to such writing, research, and thinking. * * * Meanwhile, the March 2014 AEJMC News included a column by Trevor Parry-Giles summarizing data on communication journals and their influence, measured in terms of controversial “impact factors.”1 More interesting to me, however, were his data on the growth in the number of communication journals from 1997, when there were thirty-six journals, to 2012, when the number doubled to seventy-two. Ignoring for the moment the question of what criteria were used to define a publication as a “communication journal,” it is clear that scholars in our field have more publication venues than ever, many of them focusing on particular narrow sub-fields, areas of interest, or even methods. With such specialization of interest areas, and growing numbers of scholars in those areas, the opportunities—and financial incentives for publishing houses—to create new journals increase (witness the fact that nearly a dozen of the twenty-eight divisions and interest groups in AEJMC alone have their own journals). I have even heard colleagues solemnly describe the “need” for a new journal because one of the major communication association (International Communication Association, National Communication Association, AEJMC, etc.) flagship journals does not publish “enough” articles in a specialty area, does not allow authors in that specialty sufficient space (word count) to do adequate work, or requires contributing authors to adhere to a style that is not friendly to a specialty. I envision a dammed-up torrent of unpublished work, just waiting to be set free! In terms of our “social contract” as scholars, on the other hand, I have also heard very junior colleagues say they will not participate in the broader scholarly process of reviewing, for example, for one of the flagship journals; instead, they opt for newer and narrower journals, reviewing for and joining the editorial boards of those journals. Some even expect to be paid to review!

Downloaded from jmq.sagepub.com by guest on December 20, 2014 Editorial Commentary 647

At Quarterly, we have had a practice of inviting only senior, established scholars to our board. We offer no monetary compensation. There is much to be said, of course, for specialized journals that have significant influence on and among those whose work is focused in a narrow area, who know its literature, theory, methods, and standards of evidence. Focus and depth signal maturity of an area of scholarship. Moreover, narrowly focused work may not compete well with research that has broad appeal. On the other hand, there is much to be said, too, for scholarship that is theorized, conducted, and presented in such a way that makes its own importance clear—that makes its own case—outside a narrow community. To paraphrase a question I wrote a decade ago in the Quarterly to encourage theory- developing articles, is there still a place for a journal that tries to serve across all areas of journalism and mass communication, “that cuts across specialties, makes connections, and encourages interdisciplinarity, and that addresses broad issues, themes, and theories in the field?”2 * * * In order to deal with the changing number and nature of scholarly, refereed journals, as well as technological advances, Quarterly’s submission process has undergone a number of changes during the last thirteen+ years that I have been fortunate to serve as editor. We have moved from receiving all submissions as hard copies (in quadrupli- cate!) via “snail mail,” to receiving e-mail attachments, to receiving only through an online submission site. That evolution has generally been accompanied by a steady increase in the total number of submissions. International submissions, in particular, have increased from 12% in 2003 to 42% in 2013. Quarterly has, in fact, since 2000 been friendly to manuscripts submitted in any style for review (with conversion to our style for final publication). We have relaxed our word length restriction: whereas we previously would not review a manuscript with main text of over five thousand words, we now routinely review submissions up to six thousand words and beyond. At the same time, we have increased the percentage of manuscripts receiving three reviews from 20% in 2004 to 82% in 2013. * * * Is there still a place for a journal that clearly tries to serve all areas of the field? One way to address that question is to look at who has chosen to submit their work to the journal. We gave a partial answer to this question in an editorial report on a survey of contributors in the spring 2014 issue.3 The same question can also be addressed in another way, on a different scale, by examining the institutional affiliation of contributors and published authors who have chosen Quarterly for submitting their work. If we look at roughly two decades (1993-2013) of submissions of manuscripts to Quarterly (see Table 1) and acceptances (see Table 2), it is clear that the journal has been an important venue for scholars at some of the most well-established and prestigious journalism and mass communication programs. While submissions came from over 130 colleges and universities from all over the world, and acceptances during the two decades represented more than fifty institutions, Table 1 uses first-author institutional affiliation

Downloaded from jmq.sagepub.com by guest on December 20, 2014 648 Journalism & Mass Communication Quarterly 91(4)

Table 1. Number of Articles Submitted to Journalism & Mass Communication Quarterly by First-Author Institutional Affiliation (1993-2013), Top Fifteen Only.

f University of Texas at Austin 130 University of Wisconsin 103 University of Missouri 99 Michigan State University 98 University of Florida 90 Indiana University 86 University of Georgia 84 University of Alabama 80 Southern Illinois University 73 University of North Carolina 64 Louisiana State University 58 University of Minnesota 56 Washington State University 56 Ohio State University 54 Penn State University 53 Total submissions, all affiliations 3,820

Table 2. Number of Articles Published in Journalism & Mass Communication Quarterly by First-Author Institutional Affiliation (1993-2013), Top Fifteen Only.

f Michigan State University 32 Indiana University 29 University of Texas at Austin 29 University of Florida 27 University of Alabama 26 University of Georgia 23 University of Wisconsin 23 University of Missouri 20 Washington State University 17 Ohio State University 16 Southern Illinois University 16 Penn State University 15 University of North Carolina 15 Louisiana State University 14 University of Washington 14 Total published, all affiliations 720 to identify the fifteen that were most “active” (in terms of submissions). Those fifteen universities accounted for about a third of all submissions; ten of the fifteen—all of which have doctoral programs—accounted for a fourth of all submissions. The University

Downloaded from jmq.sagepub.com by guest on December 20, 2014 Editorial Commentary 649 of Texas at Austin was most active, with 130 submissions, followed by Wisconsin, Missouri, and Michigan State. For the most part, scholarly “activity” as submissions is related to scholarly “productivity” in terms of acceptances and publications, though Table 2 shows that the fifteen most productive programs accounted for 44% of acceptances (compared to a third of submissions), and the top ten most productive programs accounted for 34% of acceptances (compared to a fourth of submissions). The ranking of the universities changes slightly when the focus is on productivity: Michigan State was the most productive, edging Indiana University and Texas–Austin. Of course, the scholarly activity and productivity data reported in the tables are based only on the affiliation of the first author as identified with the submission of the manuscript, are limited to a twenty-one-year period, and represent only the submissions and acceptances of a single journal, Journalism & Mass Communication Quarterly, which will soon enter its tenth decade of publication. * * * When I began my service as Quarterly editor with the spring 2002 issue, I noted that there are times when an editor of a journal is particularly aware of the contributions others make to publishing the journal. One of those times is at the end of the term, looking back, when an editor cannot help but reflect on the hundreds of manuscripts, thousands of reviews, and hundreds of edited proofs that have crossed the desk. Thanks are due to AEJMC, its Publications Committee, and its administrative staff; to former editors Guido H. Stempel III, Donald L. Shaw, and Jean Folkerts; to the associate and book section editors who have been part of the team; to the members of the Editorial Advisory Board; to hundreds of ad hoc reviewers; and to the Scripps School of Journalism at Ohio and the School of Journalism and Mass Communication at University of North Carolina at Chapel Hill. Obviously, I am also grateful to the hundreds—perhaps thousands—of authors who continue to believe that, yes, Quarterly is the place for their work. I owe special gratitude to friend and Production Manager Lillian Coleman, and to friend and Editorial Assistant Nancy Pawlow. Finally, thank you to DeeDee, who shared me with the journal all these years.

Daniel Riffe Editor Richard Cole Eminent Professor Jesse Abdenour, Data Analysis Park PhD Fellow UNC-Chapel Hill * * * Congratulations to the next editor of Journalism & Mass Communication Quarterly, Dr. Louisa Ha, a professor in the School of Media and Communication at Bowling Green State University. Dr. Ha has served as a member of the Editorial Advisory Board and was named associate editor in 2007. She may be reached at [email protected]. Note: Among other initiatives she has proposed, Dr. Ha will be considering submissions of more than

Downloaded from jmq.sagepub.com by guest on December 20, 2014 650 Journalism & Mass Communication Quarterly 91(4)

6,500 words, and she will be overseeing Quarterly’s move to APA (American Psychological Association) style for final publication of articles. For guidance, see the revised Instructions for Contributors in this issue or online at http://www.sagepub.com/ journals/Journal202061/manuscriptSubmission#tabview=manuscriptSubmission

Notes 1. Trevor Parry-Giles, “Journal Impact Factors and Communication Journals: A Report from the National Communication Association,” AEJMC News 47 (March 2014): 8-9. 2. Daniel Riffe, “An Editorial Comment,” Journalism & Mass Communication Quarterly 82 (spring 2005): 2-3. 3. Brendan R. Watson and Daniel Riffe, “Who Submits Work to JMCQ and Why? A Demographic Profile and Belief Summary,” Journalism & Mass Communication Quarterly 91 (spring 2014): 5-16.

Downloaded from jmq.sagepub.com by guest on December 20, 2014 JMQXXX10.1177/1077699014543380Journalism & Mass Communication QuarterlyHollander 543380research-article2014

Audience Motivations and Responses

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 651–668 The Surprised Loser: The © 2014 AEJMC Reprints and permissions: Role of Electoral Expectations sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014543380 and News Media Exposure in jmcq.sagepub.com Satisfaction with Democracy

Barry A. Hollander1

Abstract People tend to predict their preferred political candidate will win an election. The resulting “surprised losers” are examined as playing a role in the established gap between electoral winners and losers on such important outcomes as trust in government and satisfaction with the electoral process and democracy. Exposure to Fox News programming can increase wishful thinking among Republican candidate supporters, but among surprised losers, it does not increase negative attitudes.

Keywords quantitative, methodology, political, communication effects, political communication, television

On Election Night 2012, the Fox News Channel projected U.S. President Barack Obama had won Ohio—and thus re-election—yet the cable news network had a problem. Karl Rove disagreed. Rove, a powerful Republican and Fox pundit, criticized the network’s call as “premature” and put the Fox news anchors in an awkward position. On one side stood a leading political strategist who insisted Republican Party nominee Mitt Romney remained a viable candidate, while on the other side stood the network’s own analysts who projected they were “99.9%” certain of the call. On air, Rove continued to disagree.1 Rove was wrong, but he was not alone. When it comes to predicting a winner, people draw heavily on their own preferences. This preference-expectation link, called

1University of Georgia, Athens, GA, USA Corresponding Author: Barry A. Hollander, College of Journalism and Mass Communication, University of Georgia, 120 Hooper Street, Athens, GA 30602, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Downloaded from http://www.elearnica.ir 652 Journalism & Mass Communication Quarterly 91(4) wishful thinking, is a robust cognitive and motivational bias found in domains ranging from sports to politics and spanning different countries.2 Put simply, people tend to predict their preferred party or candidate will be victorious, a biased perception so stubborn it often continues even in the face of correcting information. The focus here is not merely on a theoretically interesting effect, but rather on its contribution to our understanding of an unrelated, but long-running, body of scholarly work that posits a thriving democracy hinges not on the electoral winners but, rather, lies “in the hands of the losers.”3 While scholars have identified differences between electoral winners and losers in their trust and satisfaction with democracy, they have been less successful in identifying the underlying individual-level causes of such gaps.4 Analyses below will examine a previously unexplored factor of losing, the surprised loser. The use of “surprised” in this instance signifies respondents who engaged in wishful thinking by expecting their losing candidate to win, as opposed to fellow supporters unsurprised by the election outcome. In addition, the role of exposure to news media, theoretically expected to improve knowledge about a campaign, is explored in two ways. First, while general news exposure should reduce wishful thinking and lead to less of a “surprise” in the electoral outcome, it is argued that exposure to partisan news media favorable to the losing candidate will actually increase this surprise and, as a consequence, result in even more negative perceptions of government and democracy.

Literature Review Winners and Losers Everyone loves a winner, so goes the old saying, and electoral losers received little scholarly attention until 1983 when Riker subjected them to one of the earliest systematic analyses. He wrote,

The dynamics of politics is in the hands of the losers. It is they who decide when and how and whether to fight on. Winners have won and do not immediately need to change things. But losers have nothing and gain nothing unless they continue to try to bring about new political situations.5

In their follow-up, comprehensive analysis across democracies, Losers’ Consent, Anderson and his colleagues echoed Riker, arguing that winning and losing are fundamental to understanding how people construct their political worlds and, more broadly, how their acceptance or rejection can influence the levels of support in the legitimacy of major democratic institutions.6 Indeed, the authors posit that “to say the functioning and the maintenance of democratic polities are intimately linked with what and how citizens think about democratic governance is perilously close to stating a tautology.”7 This view flows naturally from the classic work of Easton, who positioned democratic legitimacy as resting on the extent to which people rely on their government to act appropriately in most instances.8 Studies in this tradition typically focus on either a macro or micro approach. At the macro level, research has found winner–loser gaps to

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 653

be a function of differences between mature and emerging democracies,9 between elections that were close versus those that were not close,10 and whether analysis includes both presidential and congressional elections.11 At the micro level, such concepts as trust in government, belief in the legitimacy of electoral outcomes, and satisfaction with democracy dominate the literature.12 In general, individuals most involved in the political system, both behaviorally and psychologically, tend to give their democratic institutions higher levels of system support.13 A host of sociodemographic and political factors, including exposure to news media, have also been identified as possibly playing a role in system support.14 Anderson and colleagues identified three theoretical explanations for the winner– loser gap.15 From a utilitarian perspective, individuals prefer winning for the obvious benefits attached to it.16 From an affective perspective, different emotional responses are seen from winning (euphoria) and losing (anger and disillusionment).17 Finally, the dominant approach is a host of cognitive explanations based primarily on dissonance theories, which rest on the notion that people are motivated to maintain consistency in their beliefs and attitudes.18 Such dissonance can prompt supporters of the losing side to adjust the view of their preferred candidate to become more similar to the winner.19

Wishful Thinking While the gap between winners and losers at the individual level is well established theoretically, little progress has been made in successfully identifying consistent predictors of the gap.20 One potential individual-level explanation offered here for this “losers’ lament” is the difference between those who expected their preferred candidate or party to lose the election versus those who expected to win and, therefore, were surprised by the result. As such, this weaves together the theoretical threads of wishful thinking and the winner–loser gap to argue the negative effects of losing on institutional support may, in part, be a function of those surprised at losing being more negative than fellow losers unsurprised by the outcome. Scholars long ago noted the tendency of people to expect their preferred candidate to win an election,21 and subsequent studies confirmed the existence of wishful thinking, conceptualized as the link between preference and expectations.22 Research has found this effect to be “reason- ably strong, reliable, and pervasive,”23 and studies from Israel,24 Sweden,25 New Zealand,26 and the United States27 supported its extension into political campaigns. When asked to make a prediction, so goes the dominant theoretical reasoning, individuals rely on a biased sample of likeminded others and the chronic accessibility of their own opinions. In a similar vein, the theory of motivated reasoning supports the notion that people choose to believe what they want to believe, often regardless of evidence to the contrary.28 Certain factors are theoretically believed to enhance wishful thinking, such as the intensity of affect toward a candidate or the electoral outcome.29 Factors believed to reduce wishful thinking, such as political knowledge, have been found to have a neg- ligible effect.30 Theoretically, news media exposure is also thought to reduce wishful

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 654 Journalism & Mass Communication Quarterly 91(4) thinking by exposing people to polling and other countervailing information, thus making them less likely to predict their losing candidate will win. However, most studies find either no influence by exposure to news,31 or, at best, modest influence.32 This may be, in part, due to the fragmented nature of news consumption and the inability of broad measures of exposure to capture preferences among cable news channels, Internet sites, and the like.

Media and Government The erosion of public trust in government is well established.33 Research in the winner–loser gap often identifies news coverage, particularly network television news coverage, as a culprit in both lowering confidence in government and negatively influencing trust and confidence in other major institutions. Indeed, studies suggest the kinds of coverage televised in Western democracies to be associated with greater cynicism and declining civic engagement, research often identified as in the “video malaise” tradition,34 and that changes in the media are contributing to a reshaping of the citizen–state relationship.35 In addition, recent work finds time spent on the Internet to also be associated with less trust in government.36 While studies of the winner–loser gap often discuss the role of the news media in lowering confidence or satisfaction with government and democracy, few include media variables in their analyses. Anderson and LoTempio examined survey and electoral data from the 1972 and 1996 U.S. presidential elections to validate the winner– loser gap on post-election government trust and found interest in public affairs, their nearest proxy of news media consumption, to have no effect on the gap.37 Banducci and Karp used panel data from the United States, the United Kingdom, and New Zealand to find small but significant changes in efficacy and trust from before and after the election, with media coverage demonstrating a generally positive effect in two of the countries, but not in the United States.38 Exposure to talk radio and other non-traditional media, typically more partisan or ideological in content, demonstrated negative effects on trust and efficacy. This latter study dovetails nicely with recent work concerning the fragmentation of the U.S. media marketplace, particularly in terms of cable news and especially studies that examine both the content and exposure to the conservative-leaning Fox News Channel. A growing body of work has found Fox News to be more supportive of Republican or conservative positions, or more critical of Democratic or liberal positions.39 Not surprisingly, there has been a migration over recent years by conservatives and Republicans to Fox News and away from other cable channels and, to a lesser degree, a shift of Democrats and liberals to MSNBC.40 This has resulted in differential effects for exposure to Fox News in particular, such as misperceiving the existence of “death panels” in the health care overhaul.41 These studies are often framed in the theoretical perspective of motivated reasoning, which rests heavily on theories of cognitive consistency, factors also thought to play a role in both wishful thinking and the electoral winner–loser gap.42 As Anderson and his colleagues note, “ideologues are particularly prone to view the system through the lens of winning and losing,” suggesting that migration to partisan sources of news

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 655

could exacerbate the winner–loser gap and, by extension, attitudes about democracy and elections.43 In 2012, a number of analysts, most famously Nate Silver of the 538 blog, accurately predicted via various aggregation models of state and national polls that Obama would win the election.44 Individual polls of varying quality, however, suggested a tighter race, although often with Obama holding a slight edge.45 Thus, this leaves room for Romney supporters to perhaps be “surprised” by an outcome. Based on the theories and established research above, in particular the threads of wishful thinking and the winner–loser gap, it is expected:

H1: The preference for a candidate will be associated with the prediction that candidate will win the election. H2: Losers will be more negative about government, elections, and democracy than will winners. H3: Losers who predicted their candidate would win the election (surprised losers) will be more negative about government, elections, and democracy than losers who predicted their candidate would lose (expected losers).

In addition, two competing hypotheses emerge to predict the role of the more generic aspects of media exposure, such as to national television news. Existing theory suggests exposure to correcting information often found in news accounts about the likely election outcome should lead to reduced wishful thinking among people who prefer the losing candidate. Recent studies on the migration trends of the audience to sources of news compatible with their own predispositions and the biased processing of such news, however, suggest exposure to specific channels of news may increase wishful thinking. While little work has been conducted on media use and wishful thinking, the results above, and theoretical reasoning, support the expectation that exposure to generic or global forms of news should, in aggregate, result in exposure to polls and other correcting information about a campaign. The likely outcome, then, should result in such exposure being associated with less wishful thinking, especially among the eventual losers in the campaign. In terms of the more partisan Fox News, however, given the nature of the 2012 election campaign in which the Republican candidate lost, and based on previous research findings, there should be an opposite effect. Therefore,

H4: Exposure to national television news, paper newspapers, and Internet news will be associated with less wishful thinking among supporters of the losing candidate, Mitt Romney. H5: Exposure to Fox News will increase the wishful thinking effect for supporters of Romney. H6: Exposure to Fox News will be associated with more negative attitudes about government, elections, and democracy. H7: Exposure to Fox News will predict more negative attitudes about government, elections, and democracy among surprised losers than expected losers.

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 656 Journalism & Mass Communication Quarterly 91(4)

Method Data were drawn from a preliminary release of the American National Election Study’s 2012 Time Series Study, which included a pre-election and post-election survey of the same U.S. voting-eligible adults.46 Of interest here is the set of questions in which respondents were asked, prior to the election, to predict its outcome.47 Analysis focused on the two major party candidates, Obama and Romney. In the pre-election survey, respondents were also asked for their preferred candidate and, in the post-election survey, who they voted for, with agreement in the combination of these two items satisfying the preference-expectation link that lies at the heart of wishful thinking. In a similar fashion, these items and the election outcome act as the operational definition of surprised losers—respondents who preferred Romney and expected him to win. Similarly, expected losers were those respondents who preferred Romney but who correctly predicted his loss. Four measures of media exposure were analyzed. The first three examined the traditional 0-7-days-a-week exposure to news via the Internet, national television, and paper newspapers.48 The data did not include an overall measure of exposure to Fox News but rather a set of dichotomous exposure items to programs found on the channel, such as The Five, The O’Reilly Factor, and Hannity.49 These were summed into an overall index of Fox News exposure. A similar list of programs for other television networks, such as CNN or MSNBC, was not available in the data. A lack of trust in the electoral process or its outcomes can erode citizens’ perceptions of the legitimacy of governmental and democratic institutions.50 Available in the data were five criterion variables often found in previous research on gaps between winners and losers, all measuring to some degree respondent attitudes toward government, elections, or democracy. The first is the perception that the election was fairly conducted. Electoral Integrity was measured by combining a set of four questions into an index. The questions asked respondents how often, on 1-4 scales, votes were counted fairly, journalists provided fair coverage of elections, election officials acted in a fair manner, and voters were offered a “genuine choice at the ballot box.” Also in the winner–loser literature is the perception that elections have positive consequences, thus electoral responsiveness was measured by a single item that asked respondents “how much do you feel having elections makes the government pay attention to what the people think?” Responses ranged on a 1-3 scale from “not much” to “a great deal.” Satisfaction with democracy was measured on a 1-4 scale with respondents asked their degree of satisfaction with “the way democracy works in the United States.” The most extreme of the measures, Government Poses a Threat, was measured by first asking respondents whether “the federal government’s powers pose a threat to the rights and freedoms of ordinary citizens, or not?” A “no” answer categorized respondents as the lowest score and a follow-up question asked the degree of the threat. The result was a 1-5 scale with the high score representing the perception of greater threat. The four previous criterion variables were measured only in the post-election survey. Trust in Government was created by a change score calculated by subtracting a single pre- election question from a single post-election question. The pre-election question

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 657

asked, on a 1-5 scale, to what degree respondents felt those “running the government” were corrupt. The post-election question asked, on a 1-3 scale, to what degree respondents felt those “running the government” were crooked. The results of each were centered and a change score computed for analysis.51

Results The pre-election and post-election surveys resulted in 2,054 completed face-to-face interviews and 3,860 Internet-based interviews from universes of U.S. eligible voters. Pre-election interviews were conducted during the two months prior to the 2012 election and were followed by post-election re-interviewing beginning November 7.52 All data in subsequent analyses were weighted to represent the population.53 Respondents in the pre-election survey demonstrated substantial wishful thinking, the link between preference and expectation of victory (χ2 = 2,823.4, df = 4, p < .001). Among Obama supporters, 97% predicted he would be victorious, while among Romney supporters, 78% predicted he would win the election, supporting H1.54 Few respondents changed their minds from the pre-election and post-election surveys. Among those who before the election preferred Obama, 98% reported having voted for him in the post-election survey, while among those who said before the election they preferred Romney, 97% reported having voted for him.55 Consistent factors in the likelihood to see one’s own candidate as the likely winner—regardless of candidate preferred—were party identification, ideology, affect toward the candidates, exposure to national television news, and exposure to Fox News programming, suggesting Fox News has an impact unique to preference.56 The strongest single predictor of the belief Romney (the eventual loser) would win the election was, not surprisingly, a pre-election preference for Romney (B = 3.2, Wald = 479.4, p < .001, see Table 1). Caring about the election outcome and strength of self-identification as a Republican also significantly contributed to the model.57 Respondents who were consistent in their candidate preference in both the pre- election and post-election surveys were used in subsequent analyses. Thus, to catego- rize respondents as an expected loser or a surprised loser required they preferred Romney before the election, reported voting for him after the election, and predicted he would win. In this coding scheme, 305 respondents were categorized as expected losers (preferred Romney but predicted he would lose) and 1,193 respondents as surprised losers (preferred Romney and predicted he would win). While few demographic differences emerged, surprised losers, compared with expected losers, were more politically conservative, more likely to be self-identified Republicans, cared more about the election outcome, and had more positive feelings toward Romney and more negative feelings toward Obama. These bivariate results support previous work in wishful thinking that suggests affect plays a vital role in the preference-expectation link.58 In addition, surprised losers, as compared with expected losers, reported greater television news exposure (t = −3.2, df = 1,494, p < .01), greater exposure to Fox News programming (t = −4.9, df = 1,496, p < .001), and less exposure to print newspapers (t = 2.2, df = 1,494, p < .05).59 No relationship was seen for Internet news exposure.60

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 658 Journalism & Mass Communication Quarterly 91(4)

Table 1. Logistic Regression on Predictions of a Romney Win in 2012 Election (All Respondents).

B SE Wald Prefer Romney 3.24 0.15 479.4*** Age −0.00 0.02 0.0 Education −0.11 0.05 4.1* Income −0.03 0.01 11.5*** Sex (female) 0.00 0.10 0.0 White 0.08 0.15 0.3 Black −0.85 0.29 8.7** Care who wins 0.83 0.16 27.5*** Party ID (Republican) 0.29 0.03 73.7*** Internet news exp −0.03 0.02 2.0 Print newspaper exp −0.05 0.02 5.8* Television news exp −0.03 0.02 1.5 Fox News exp 0.21 0.04 25.3*** −2 Log likelihood 2,567.1 Cox & Snell R2 .49

Note. Entries are unstandardized logistic regression coefficients. Weighted data. *p < .05. **p < .01. ***p < .001.

H2 predicted election losers would be more negative about government and democracy, as compared with winners. As discussed earlier, four of the five criterion variables are positive in nature (satisfaction with democracy, change in trust in government, electoral responsiveness, and electoral integrity), while one is negative (government poses a threat).61 Some support is seen for this hypothesis, at least at the bivariate level (Table 2). Both expected and surprised losers were lower than winners in satisfaction with democracy, and perceptions of electoral integrity, and more likely to perceive government as posing a threat. In addition, surprised losers were lower than winners on perceptions of electoral responsiveness. H3 predicted surprised losers would be more negative about government and democracy than expected losers. The bivariate results provide less support for this hypothesis. Of the five criterion variables, surprised losers were more likely than expected losers to see government posing a threat and to perceive less electoral integrity. Table 3 provides the results of a more rigorous test of these hypotheses. A set of sociodemographic and political variables were regressed against the five criterion variables and then a block entered that included expected losers and surprised losers, with winners as the contrast group. The model is completed with the inclusion of a block of news media exposure variables. As Table 3 shows, the two loser categories, as contrasted with “winners,” were associated with negative opinions about democracy and government even after controlling for other factors, offering further support of H2. The inclusion of the two loser categories added modestly, but significantly, to

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 659

Table 2. Comparison of Expected and Surprised Losers and Winners by Key Variables.

Expected losers Surprised losers Election winners Internet news exposure 3.8 3.8 3.8 Television news exposure 3.8a 4.4abc 4.0bc Print newspaper exposure 2.8ab 2.4a 2.3b Fox News exposure 0.8ab 1.4ac 0.2bc Satisfied with democracy 2.5a 2.6b 2.9ab Electoral responsiveness 2.1 2.1a 2.2a Government poses threat 2.4ab 3.1ac 1.7bc Electoral integrity 2.8ab 2.6ac 3.0bc Change in Govt. trust −0.15 −0.17 −0.16

Note. Row entries are mean scores with those sharing a superscript significantly different at .05 level by t test. Weighted data.

the multivariate models, with R2 ranging from .01 to .05, depending on the criterion variable examined. Specifically, both expected losers and surprised losers, as contrasted to winners, were less satisfied with democracy (β = −.10, p < .001, in both cases). Similarly, on the pre-election and post-election trust in government change score, expected losers (β = −.05, p < .05) and surprised losers (β = −.12, p < .001) became more negative as compared with electoral winners.62 The lone exception was with perceptions of electoral responsiveness in which losers, regardless of anticipated electoral outcome, did not differ from winners. Less consistent support was seen for H3, which predicted surprised losers would be more negative about government than expected losers. No difference was seen between the two loser categorizes in satisfaction with democracy, trust in government, or perceived electoral responsiveness. However, surprised losers were more likely than expected losers to perceive the government as a threat (t = 30.4, df = 2,928, p < .001), and to question the integrity of the election (t = −19.3, df = 2,976, p < .001). The next set of hypotheses focuses on the role of media exposure. H4 predicted that exposure to generic print, television, and Internet news would result in less wishful thinking even among respondents who preferred the eventual losing candidate, while H5 predicted an opposite effect for exposure to programming featured on the conservative cable news network, Fox News. The underlying theoretical explanation is that exposure to news as measured by such global constructs as television news will likely result in exposure to more correcting information via polls and other information, therefore reducing wishful thinking among Romney supporters. The opposite is hypothesized for exposure to Fox News programming. Table 4 provides the results of a hierarchical logistic regression conducted only among Romney supporters on the perception Romney would win. Among these respondents, the most powerful predictors of wishful thinking were affect toward the election and the strength of identification as a Republican. Support for H4 is found only in the case of print newspaper exposure (B = −0.08, Wald = 8.4, p < .01) and not the other two global measures. That

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 660 Journalism & Mass Communication Quarterly 91(4)

Table 3. Hierarchical Regression on Criterion Variables by Media Exposure and Winner– Loser Status.

Satisfied with Electoral Government Electoral Change in trust democracy responsive poses threat integrity in government Age .04 −.05* −.03 .03 −.03 Education .01 .07*** −.07*** .06*** −.04 Income .07*** .01 −.08*** .07*** .00 Sex (female) −.03 .01 −.04* −.05** −.02 White −.06** −.07** −.03 .07*** −.02 Black .03 .04 −.05** .05* −.04 R2 change .02*** .02*** .05*** .02*** .01* Care about .01 .07*** .03 .01 −.01 election Party ID −.08** −.03 .08** .01 −.06* (Republican) R2 change .03*** .01*** .15*** .06*** .00* Expected loser −.10*** .03 .11*** −.13*** .03 Surprised loser −.10*** −.04 .36*** −.33*** .02 R2 change .01*** .00 .05*** .04*** .00 Internet news −.05** −.02 .03* −.03 .01 exposure TV news .05* .07** −.09*** .03 −.05* exposure Newspaper .03 .02 .03 .07*** .01 exposure Fox News −.03 .11*** .16*** −.07*** −.01 exposure R2 change .01** .02*** .02*** .01*** .00 Total R2 .06*** .04*** .27*** .13*** .01**

Note. Entries are standardized regression coefficients (beta weights), with weighted data. For white and black, contrast group is Hispanic. For expected and surprised loser, the contrast group is winners. Only final models shown. *p < .05. **p < .01. ***p < .001. is, exposure to traditional print newspapers, even after controlling for other factors, resulted in Romney supporters being less likely to predict their candidate would win the election. Support for H5, concerning Fox News, was also seen (B = 0.13, Wald = 7.2, p < .01). In other words, after sociodemographic and political controls, exposure to Fox News increased among Romney supporters the perception he would win. H6 predicted exposure to Fox News programming would predict more negative attitudes about government, while H7 extended this by predicting an effect strongest among surprised losers. Modest support was found for H6. Exposure to Fox News programming was unrelated in the multivariate models to satisfaction with democracy and trust in government, and it had a positive rather than negative association with

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 661

Table 4. Logistic Regression on Wishful Thinking (Only Romney Supporters).

B SE Wald Age −0.02 0.03 0.3 Education −0.08 0.07 1.2 Income −0.03 0.01 7.6** Sex (Female) 0.07 0.14 0.2 White −0.46 3.01 0.3 Black −1.39 0.65 4.5* Care who wins 1.25 0.22 32.0*** Party ID (Republican) 0.21 0.05 18.7*** Internet news exp −0.02 0.03 0.7 Print newspaper exp −0.08 0.03 8.4** Television news exp 0.04 0.02 1.7 Fox News exp 0.13 0.05 7.2** −2 Log likelihood 1,341.2 Cox & Snell R2 .07

Note. Entries are unstandardized regression coefficients. Weighted data. *p < .05. **p < .01. ***p < .001. perceived electoral responsiveness. However, exposure to Fox News, even after statistical controls, predicted a greater perception of the government posing a threat (β = .16, p < .001) and less perceived electoral integrity (β = .07, p < .001). In a similar vein, little support was found for H7. Analysis found few differences in the predictive power of Fox News exposure regardless of being an expected loser or a surprised loser on any of the five criterion variables. For example, Fox News predicted the perception of government as a threat for both surprised losers (β = .22, p < .001) and expected losers (β = .20, p < .01). Thus, H7 was not supported.63

Discussion When it comes to an electoral outcome, “what is key,” according to Anderson and his colleagues, “is how people react to loss; in particular, how rebellious or how apathetic a reaction is invoked.”64 Political theorists consider the acceptance and consent of election losers to be among the necessary ingredients of a successful democracy. Without losers’ consent, new policy initiatives become problematic and democracies risk further instability. Scholars have identified key gaps between winners and losers after an election; simply put, the winners are happy, the losers less so, and it is the latter that have drawn scrutiny in trying to understand under what conditions these gaps occur and the subsequent consequences in how people view their government, the electoral process, and democracy. Despite explanatory successes at the macro level, such as the closeness of an election and the maturity of a nation’s democracy, the underlying explanations at

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 662 Journalism & Mass Communication Quarterly 91(4) the individual level remain open to debate. This study examined electoral outcome expectation as an explanatory factor. Previous work in wishful thinking has long established that people tend to predict their preferred candidate will win an election, even when that candidate is losing. Here, respondents were categorized as winners (supported the eventual victor, Barack Obama) or as either expected losers (predicted their candidate, Mitt Romney, would lose) or surprised losers (predicted their candidate would win). Losers are expected to be more negative about an election, but the argument here is surprised losers will demonstrate even more doubt as to the legitimacy of the electoral process, the responsiveness of those elected, and in general reveal less trust in democracy and government. Tests of this supposition, based in part on consistency theories, received modest support, enough to warrant further study perhaps, but not so strong as to suggest the surprised loser explains the negative perceptions seen in previous studies of the winner–loser gap. As such, the search for useful individual-level explanations for the winner–loser gap does not end here. The strongest differences occur among the more strident measure presented in the post-election national survey data, the perception that government poses a threat to its citizens. Here, the surprised losers score significantly higher than expected losers, suggesting that in the most affective of attitudes about government and democracy, the kind of loser one is—surprised versus expected—may indeed play a role. A related analysis examined the role of partisan and generic media exposure in the likelihood of predicting one’s own candidate would be the winner, and the subsequent association with attitudes about democracy and the election. Previous theory and research support the notion that, in general, exposure to the news media should lead to greater knowledge and, by extension, less wishful thinking. Recent research and content analyses of Fox News, however, suggest the partisan nature of such programming could result in greater wishful thinking among those who supported the eventual loser, Romney. While the models reveal exposure to Fox News Channel’s programming tends to predict some negative attitudes, even after a number of statistical controls, such exposure did not, as hypothesized, demonstrate greater predictive power among those categorized as surprised losers. It may be that a ceiling effect occurred. Given that exposure to the conservative Fox News leads to a greater likelihood of being a surprised loser, it is hardly surprising that the combination of the two has little unique variance left to explain in the models. As expected, analysis revealed differences between traditional measures of global news exposure, such as television news exposure, and a more specific measure of exposure to programming featured by the more conservative-leaning Fox News. Exposure to programs featured on Fox News, such as those hosted by Bill O’Reilly and Sean Hannity, resulted in a greater wishful thinking effect by Romney supporters. In other words, while Romney supporters were substantially more likely to predict their candidate would win the 2012 presidential election, watching Fox News programming exacerbated this effect. As hypothesized, exposure to news via paper newspapers (but not the Internet or network television news) reduced wishful thinking among Romney supporters. The Fox News and global measure results, taken together,

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 663 demonstrate again the power of exposure to likeminded sources of information in increasing the perception that people—or an election result—agree with one’s own expectations. That at least one global media measure reduced the likelihood of this misperception suggests that these general measures capture exposure to correcting information, at least in the case of Romney supporters in the 2012 election. The lack of an effect by the global television exposure measure, however, suggests this broad measure captures some of the migration to partisan-leaning sources seen in other studies and may need to be reconsidered in further research and, when possible, replaced by more network- or program-specific measures. One strength in the analysis above is its use of a national sample of respondents interviewed before and after the 2012 election, but the results are limited by the nature of the available media variables, some single-item measures, and, importantly, the number of respondents who fell into the expected loser category—those who preferred Romney but did not predict he would win. The standard global exposure items are noted for their measurement error and inflated scores, and yet much can be learned when comparing the results with those of network- or even program-specific measures of exposure or attention.65 In addition, the large number of respondents surveyed can provide statistically significant results that are, in and of themselves, relatively small. These surprised losers did demonstrate higher scores on television news exposure and Fox News exposure, while print exposure is more associated with being an expected loser—suggesting, then, differences in the sources of information when it comes to how one perceives the outcome of an election. Regardless, the results provide a potential avenue to understand the winner–loser gap as well as introduce the role of news media exposure, in particular partisan news media, as a factor in making respondents more likely to perceive their candidate as being a winner, which could result in more negative attitudes about elections and democracy when the election outcome turns out a surprise.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. A search on YouTube will result in thousands of hits on the set of terms, “Karl Rove Election Night,” and the site prompts the user to add the term meltdown. One link is https://www.youtube.com/watch?v=9TwuR0jCavk (accessed October 23, 2013). 2. Elisha Babad and Eitan Yacobos, “Wish and Reality in Voters’ Predictions of Election Outcomes,” Political Psychology 14 (1, 1993): 37-54; Elisha Babad, “Can Accurate Knowledge Reduce Wishful Thinking in Voters’ Predictions of Election Outcomes?” The Journal of Psychology 129 (3, 1995): 285-300.

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 664 Journalism & Mass Communication Quarterly 91(4)

3. William Riker, “Political Theory and the Art of Heresthetics,” in Political Science: The State of the Discipline, ed. Ada W. Finifter (Washington, DC: American Political Science Association, 1983), 62. 4. One exception is the comparative work that finds individuals in more established democracies less likely than those in emerging democracies to demonstrate such an effect. 5. Riker, “Political Theory,” 62. 6. Christopher J. Anderson, Andre Blais, Shaun Bowler, Todd Donovan, and Ola Listhaug, Losers’ Consent: Elections and Democratic Legitimacy (Oxford, UK: Oxford University Press, 2005). See also Richard Nadeau and Andre Blais, “Accepting the Election Outcome: The Effect on Participation on Losers’ Consent,” British Journal of Political Science 23 (4, 1993): 553-63. 7. Anderson et al., Losers’ Consent, 2. 8. David Easton, A Systems Analysis of Political Life (NY: Wiley, 1965). 9. Anderson et al., Losers’ Consent. 10. Patrick Howell and Florian Justwan, “Nail-Biters and No-Contests: The Effect of Electoral Margins on Satisfaction with Democracy in Winners and Losers,” Electoral Studies 32 (June 2013): 334-43; Shane Singh, Ignacio Lago, and Andre Blais, “Winning and Competitiveness as Determinants of Political Support,” Social Science Quarterly 92 (September 2011): 695-709. 11. Christopher J. Anderson and Andrew LoTempio, “Winning, Losing and Political Trust in America,” British Journal of Political Science 32 (April 2002): 335-51. 12. Here, I focus exclusively on individual-level explanations, not comparatively across democracies that differ in terms of maturity or in other, broad system-based explanations. 13. Nadeau and Blais, Accepting; Steven Finkel, “Reciprocal Effects of Participation and Political Efficacy: A Panel Analysis,” American Journal of Political Science 29 (4, 1985): 891-913; Christopher J. Anderson and Christine A. Guillory, “Political Institutions and Satisfaction with Democracy,” American Political Science Review 91 (1, 1997): 66-81; Gabriel A. Almond and Sidney Verba, The Civic Culture (Princeton, NJ: Princeton University Press, 1963); Benjamin Ginsberg and Robert Weisberg, “Elections and the Mobilization of Popular Support,” American Journal of Political Science 22 (1, 1978): 31-55; see Anderson et al., Losers’ Consent, for a discussion. 14. Among these are age, income, voting, and education, though the latter has come under attack of late. Almond and Verba, The Civic Culture; Ola Listhaug and Matti Wiberg, “Confidence in Political Institutions,” in Citizens and the State, ed. Hans-Dieter Klingemann and Dieter Fuchs (NY: Oxford University Press, 1995), 298-322; Russell J. Dalton, Democratic Challenges, Democratic Choices: The Erosion in Political Support in Advanced Industrial Democracies (Oxford, UK: Oxford University Press, 2004). 15. Leon Festinger, A Theory of Cognitive Dissonance (Stanford, CA: Stanford University Press, 1957); Donald Granberg, “Political Perception,” in Explorations in Political Psychology, ed. Shanto Iyengar and William J. McGuire (Durham, NC: Duke University Press, 1993), 70-112. 16. Riker, “Political Theory”; Amos Tversky and Daniel Kahneman, “Advances in Prospect Theory: Cumulative Representation of Uncertainty,” Journal of Risk and Uncertainty 5 (4, 1992): 297-323. 17. Soren Holmberg, “Down and Down We Go: Political Trust in Sweden,” in Critical Citizens, ed. Pippa Norris (NY: Oxford University Press, 1999), 103-122. 18. Festinger, A Theory; William McGuire, “Theory of the Structure of Human Thought,” in Theories of Cognitive Consistency: A Sourcebook, ed. Robert P. Abelson (Chicago, IL: Rand McNally, 1968), 140-162.

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 665

19. Jack W. Brehm, “Postdecision Changes in the Desirability of Alternatives,” Journal of Abnormal Social Psychology 52 (3, 1956): 384-89; Jack W. Brehm, “Motivational Effects of Cognitive Dissonance,” in Explorations in Cognitive Dissonance, ed., Jack W. Brehm and Arthur R. Cohen (NY: John Wiley, 1962), 131-159; Allan J. Cigler and Russell Getter, “Conflict Reduction in the Post-election Period: A Test of the Depolarization Thesis,” Western Political Quarterly 30 (3, 1977): 363-76. 20. Stephen C. Craig, Michael D. Martinez, Jason Gainous, and James G. Kane, “Winners, Losers, and Election Context: Voter Responses to the 2000 Presidential Election,” Political Research Quarterly 59 (December 2006): 579-92. 21. Paul F. Lazarsfeld, Bernard Berelson, and Hazel Gaudet, The People’s Choice: How the Voter Makes up His Mind in a Presidential Campaign (NY: Columbia University Press, 1944). 22. Oden Frenkel and Anthony Doob, “Post-decision Dissonance at the Polling Booth,” Canadian Journal of Behavioral Science 8 (October 1976): 347-50; James B. Lemert, “Picking the Winners: Politician vs. Voter Predictions of Two Controversial Ballot Measures,” Public Opinion Quarterly 50 (1, 1986): 208-21; Dennis T. Regan and Martin Kilduff, “Optimism about Elections: Dissonance Reduction at the Ballot Box,” Political Psychology 9 (1, 1988): 101-107. 23. Donald Granberg and Soren Holmberg, The Political System Matters: Social Psychology and Voting Behavior in Sweden and the United States (Cambridge, UK: Cambridge University Press, 1988), 164. 24. Babad and Yacobos, “Wish and Reality”; Babad, “Can Accurate.” 25. Granberg and Holmberg, The Political System; Donald Granberg, “Preference, Expectations, and Placement Judgments: Some Evidence from Sweden,” Social Psychology Quarterly 46 (4, 1983): 363-68. 26. Elisha Babad, Michael Hills, and Michael O’Driscoll, “Factors Influencing Wishful Thinking and Predictions of Election Outcomes,” Basic and Applied Social Psychology 13 (4, 1992): 461-76. 27. Donald Granberg and Edward Brent, “When Prophecy Bends: The Preference-Expectation Link in U.S. Presidential Elections,” Journal of Personality and Social Psychology 45 (September 1983): 477-91; Kathleen A. Dolan and Thomas M. Holbrook, “Knowing versus Caring: The Role of Affect and Cognition in Political Perceptions,” Political Psychology 22 (1, 2001): 27-44; Barry A. Hollander, “People Think like Me: Religion and Wishful Thinking in the 2000 U.S. Presidential Election,” Journal of Media and Religion 3 (4, 2004): 187-97. 28. Ziva Kunda, “The Case for Motivated Reasoning,” Psychological Bulletin 108 (November 1990): 480-98. 29. Granberg and Holmberg, The Political System; Babad, Hills, and O’Driscoll, “Factors.” 30. Babad, “Can Accurate Knowledge”; Babad, Hills, and O’Driscoll, “Factors”; but see also Dolan and Holbrook, “Knowing versus Caring.” 31. Dolan and Holbrook, “Knowing versus Caring.” 32. Hollander, “People Think like Me.” 33. Paul R. Abramson and Robert Inglehart, Value Change in Global Perspective (Ann Arbor: University of Michigan Press, 1995); Russell J. Dalton, Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies, 3rd ed. (NY: Chatham House Press, 2002). 34. Robert D. Putnam, “Tuning In, Tuning Out: The Strange Disappearance of Social Capital in America,” Political Science & Politics 28 (4, 1995): 664-83; Michael J. Robinson,

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 666 Journalism & Mass Communication Quarterly 91(4)

“Public Affairs Television and the Growth of Political Malaise: The Case of the ‘The Selling of the Pentagon,’” American Political Science Review 70 (June 1976): 409-32. For a counter argument and evidence, see Eric M. Uslaner, “Social Capital, Television, and the ‘Mean World’: Trust, Optimism, and Civic Participation,” Political Psychology 19 (September 1998): 441-67; Thomas E. Patterson, Out of Order (NY: Alfred A. Knopf, 1993); Joseph N. Cappella and Kathleen H. Jamieson, Spiral of Cynicism: The Press and the Public Good (NY: Oxford University Press, 1997); Stephen Earl Bennett, Staci L. Rhine, Richard S. Flickinger, and Linda L. M. Bennett, “‘Video Malaise’ Revisited: Public Trust in the Media and Government,” Harvard International Journal of Press/Politics 4 (4, 1999): 8-23; Patricia Moy and Dietram A. Scheufele, “Media Effects on Political and Social Trust,” Journalism & Mass Communication Quarterly 77 (December 2000): 744- 59; Robinson, “Public Affairs Television and the Growth of Political Malaise”; and Arthur H. Miller, Edie N. Goldenberg, and Lutz Erbring, “Type-Set Politics: Impact of Newspaper on Public Confidence,” American Political Science Review 73 (March 1979): 67-84. 35. Margaret T. Gordon, “Public Trust in Government: The US Media as an Agent of Accountability?” International Review of Administrative Sciences 66 (2, 2000): 297-310. 36. Tobin Im, Wonhyuk Cho, Greg Porumbescu, and Jungho Park, “Internet, Trust in Government, and Citizen Compliance,” Journal of Public Administration Research and Theory 22 (4, 2012): 1-23. 37. Anderson and LoTempio, “Winning.” 38. Susan A. Banducci and Jeffrey A. Karp, “How Elections Change the Way Citizens View the Political System: Campaigns, Media Effects and Electoral Outcomes in Comparative Perspective,” British Journal of Political Science 33 (July 2003): 443-67. 39. Sean Aday, “Chasing the Bad News: An Analysis of 2005 Iraq and Afghanistan War Coverage on NBC and Fox News Channel,” Journal of Communication 60 (March 2010): 144-64; Sean Aday, Steven Livingston, and Maeve Hebert, “Embedding the Truth: A Cross-Cultural Analysis of Objectivity and Television Coverage of the Iraq War,” Harvard International Journal of Press/Politics 10 (winter 2005): 3-21; Mark D. Harmon and Robert Muenchen, “Semantic Framing of the Iraq War: Fox v. CNN and Other U.S. Broadcast News Programs,” ETC: A Review of General Semantics 66 (April 2009): 12-26; Lauren Feldman, Edward W. Maibach, Connie Roser-Renouf, and Anthony Leiserowitz, “Climate on Cable: The Nature and Impact of Global Warming Coverage on Fox News, CNN, and MSNBC,” International Journal of Press/Politics 17 (January 2012): 3-31; Matthew A. Baum and Tim Groeling, “New Media and the Polarization of American Political Discourse,” Political Communication 25 (October 2008): 345-65; Jennifer Brubaker and Gary Hanson, “The Effect of Fox News and CNN’s Postdebate Commentator Analysis on Viewers’ Perceptions of Presidential Candidate Performance,” Southern Communication Journal 74 (October 2009): 339-51; Jonathan S. Morris and Peter L. Francia, “Cable News, Public Opinion, and the 2004 Party Conventions,” Political Research Quarterly 63 (December 2010): 834-49. 40. Barry A. Hollander, “Tuning Out or Tuning Elsewhere? Partisanship, Polarization, and Media Migration from 1998 to 2006,” Journalism & Mass Communication Quarterly 85 (March 2008): 23-40; Kevin Coe, David Tewksbury, Bradley J. Bond, Kristin L. Drogos, Robert W. Porter, Ashley Yahn, and Yuanyuan Zhang, “Hostile News: Partisan Use and Perceptions of Cable News Programming,” Journal of Communication 58 (June 2008): 201-19; Shanto Iyengar and Kyu S. Hahn, “Red Media, Blue Media: Evidence of Ideological Selectivity in Media Use,” Journal of Communication 59 (March 2009): 19-39.

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 Hollander 667

41. Patrick C. Meirick, “Motivated Misperception? Party, Education, Partisan News, and Belief in ‘Death Panels,’” Journalism & Mass Communication Quarterly 90 (spring 2013): 39-57. 42. Kunda, “The Case for Wishful Thinking.” 43. Anderson et al., Losers’ Consent, 89. 44. See, for example, http://mashable.com/2012/11/07/nate-silver-wins/ (accessed March 21, 2014). 45. A Fox news poll published in late October, for example, had Obama and Romney locked at 46% apiece. Earlier in the campaign, when many respondents were surveyed, polls also had the race either within the margin of error or one candidate slightly ahead. See, for example, http://www.politico.com/p/2012-election/polls/president/national (accessed March 21, 2014). 46. Details about the ANES and its methodology are available at http://electionstudies.org/ (accessed June 30, 2014). The 2012 survey was of 2,054 completed face-to-face and 3,860 completed Internet-based instruments conducted before and after the election. 47. Respondents were also asked to predict the outcome and closeness of the presidential election in their home state. The state questions are not used here. 48. The Internet measure is, of course, unnecessarily broad and imperfect given the wide range of sources available. It is offered here as a global measure of how people get news from the medium and will reflect the wide range of sources. 49. Unfortunately, the data did not provide similar measures for other cable news networks, such as MSNBC or CNN. 50. Sarah Birch, “Perceptions of Electoral Fairness and Voter Turnout,” Comparative Political Studies 43 (December 2010): 1601-22. 51. The two items were highly correlated (r = .45, p < .001). Because they were measured on different metrics, a “centered” score for each was computed and a change score derived by subtracting the pre-election score from the post-election score. 52. Additional methodological details are available from the 2012 Time Series Study page: http://www.electionstudies.org/studypages/anes_timeseries_2012/anes_timeseries_2012. htm (accessed June 30, 2014). 53. The variable for weighting the full sample is WEIGHT_FULL. 54. Among those who preferred an “other” candidate, the effect disappears, with 57.2% of these predicting an Obama win and 27.0% predicting a Romney victory. In all analyses below, the data were weighted to be representative of the population. 55. χ2 = 4,809.0, df = 4, p < .001. Of the 138 respondents with an “other” preference before the election, 70 (50.7%) reported having voted for another candidate, while the rest almost evenly split between Obama and Romney. 56. Differences emerge if you examine wishful thinking specifically among Romney or Obama supporters. For example, among those who preferred Obama, being black resulted in greater wishful thinking. Among Romney supporters, age, affect about the campaign in general, and television news exposure all led to a greater wishful thinking effect. 57. In this case, party identification is a traditional 7-point scale with low representing Democratic self-identification and high representing Republican self-identification. 58. A multivariate logistic model to predict being a surprised loser as compared with an expected loser resulted in few variables retaining their statistical significance. Income predicted less likelihood to be surprised while affect toward Romney, being conservative, and television news exposure predicted a greater likelihood of being surprised.

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 668 Journalism & Mass Communication Quarterly 91(4)

59. The three generic news exposure items had ranges from 0 to 7. The means and standard deviations for each were Internet News Exposure (M = 3.6, SD = 2.8), Television News Exposure (M = 3.9, SD = 2.6), and Print Newspaper Exposure (M = 2.1, SD = 2.6). For the Fox News programming, the range was from 0 programs to 6. Most respondents reported no exposure (78.2%), with M = 0.5, SD = 1.3. 60. A multivariate logistic model with sociodemographic factors, as well as party identification and affect about the election outcome as covariates, still found Fox News exposure to be a significant predictor of being a surprised loser (B = 0.19, Wald = 5.4, p < .05). 61. Zero-order correlations ranged in strength from .11 to .33, either positive or negative in direction, all of them significant at the .001 level. A full description of the relationships between this set of criterion variables and the sociodemographic, political, and media variables is beyond the scope of the discussion here. Internet news exposure was associated only with the perception that the government is a threat. Television news exposure was also associated with the perception of a threat and also with two other positive perceptions (satisfaction with democracy and electoral responsiveness). Newspaper exposure was associated only with electoral integrity. Finally, exposure to Fox News programming was associated with a greater perception of government threat and in a negative fashion with all of the remaining criterion variables. A full set of correlations is available from the author. 62. The overall change from pre-election to post-election was a move in the negative direction, −0.15. Of respondents, 82% reported lower trust after the election as compared with before the election. 63. Another approach, including interaction terms in the models, also did not support H7. 64. Anderson et al., Losers’ Consent, 6. 65. For a recent treatment of these global exposure measures versus exposure to more specific network and programs, see Barry A. Hollander, “The Role of Media Use in the Recall versus Recognition of Political Knowledge,” Journal of Broadcasting & Electronic Media 58 (1, 2014): 97-113.

Downloaded from jmq.sagepub.com at b-on: 01100 Universidade do Porto on December 20, 2014 JMQXXX10.1177/1077699014550090Journalism & Mass Communication QuarterlyVu et al. 550090research-article2014

Audience Motivations and Responses

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 669–686 Exploring “the World © 2014 AEJMC Reprints and permissions: Outside and the Pictures sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550090 in Our Heads”: A Network jmcq.sagepub.com Agenda-Setting Study

Hong Tien Vu1, Lei Guo2, and Maxwell E. McCombs1

Abstract This study examines the Network Agenda Setting Model, the third level of agenda- setting theory. It seeks to expand the model’s scope by testing five years (2007-2011) of aggregated data from national news media and polls. The study finds evidence that the news media bundled issue objects and made them salient in the public’s mind. Findings of the study also demonstrate strong network correlations of issue salience among different types of news media.

Keywords network agenda setting, network analysis, intermedia agenda setting

The proposition of the Network Agenda Setting Model (NAS)1 suggested a great change in how we think about media effects. Traditional agenda-setting research assumes that issues are transferred discretely as individual elements from the news media to the public. The NAS model, however, theorizes that objects and attributes can also be transferred simultaneously in bundles between the agendas. Thus, as Guo2 argued, not only can the news media tell us what to think about and how to think about it, they are also capable of telling us what and how to associate. Importantly, the NAS model has expanded the agenda-setting theory to the third level by suggesting a distinctive theoretical perspective, that the salience of interrelationships between

1University of Texas at Austin, Austin, TX, USA 2Boston University, Boston, MA, USA Corresponding Author: Hong Tien Vu, School of Journalism, University of Texas at Austin, 300 W. Dean Keeton, Austin, TX 78712, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Downloaded from http://www.elearnica.ir 670 Journalism & Mass Communication Quarterly 91(4) issues and/or attributes can be transferred from the news media to the public. This is a much broader and more comprehensive hypothesis about media effects than traditional agenda-setting hypotheses which focused on the salience of discrete objects and attributes. In terms of agenda-setting effects, this focus on bundles of message elements provides a more nuanced understanding of the “pictures in our heads.”3 In addition, the model also provides a new methodological approach,4 which employs social network analysis to examine network relationships among objects and/or attributes, respectively, on the media and public agenda, and then compares the two. To date, two empirical studies based on data collected from Texas elections in 2002 and 2010 have been conducted to test the NAS model.5 These studies found that the interrelationships among political candidates’ attributes emphasized by the news media had significant links with the public’s perceptions (correlations of .84 and .71). The two studies demonstrated strong NAS effects, but both of them examined attributes only, and were conducted against the backdrop of the state of Texas, a local setting. Whether the NAS model could apply to object agendas and whether the model could be expanded to test data in a national setting remain unknown. To address these gaps, the present study seeks to test the NAS model through examining object-based media agenda networks and public agenda networks with nationwide aggregated data. Specifically, secondary analyses were conducted on the content analysis data initially collected by the Project for Excellence in Journalism (PEJ) and survey data collected by the Gallup polls during a five-year period: 2007 to 2011. Furthermore, this study also investigates intermedia network agenda-setting relationships between different news outlets.

Theoretical Background Network Agenda Setting Model Agenda-setting theory was introduced over forty years ago in the seminal Chapel Hill study, in which McCombs and Shaw6 proposed that the salience of issues emphasized by the news media could be transferred to the public’s mind. Since then, hundreds of empirical studies have been conducted to test the theory, cumulatively demonstrating the theory’s predictive power in various social contexts.7 In performing the first level of agenda-setting research, scholars identify the media agenda by systematically analyzing news content. To measure the public agenda, public surveys are usually conducted to ask questions such as “What do you think is the most important problem facing this country today?” Researchers then compare the rank order of objects on the media and the public agendas to determine the degree to which the two agendas are correlated. Agenda-setting research advanced to its second level by focusing on the attributes of issues and other objects.8 Methodologically, scholars compare attribute salience between the media and public agendas. Theoretically, while first-level agenda-setting theory suggests news media can affect what we think about, the second level further indicates that the media can also influence how we think about objects.9

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 671

As Guo and McCombs10 observed, first- and second-level agenda-setting research share two long-standing assumptions. The first assumption holds that humans’ mental representation operates primarily in a logical and linear model. The second assumption implies that the transfer of issue or attribute salience occurs discretely. Differing from what the first assumption suggests, studies have demonstrated that humans’ cognitive representation does not necessarily operate linearly in learning or perceiving the world; rather, it could be presented as a network-like structure.11 Scholars thus proposed concepts such as “associative network model,”12 “cognitive mapping,”13 and “cognitive network model”14 to illustrate cognitive representation. Guo and McCombs15 borrowed these concepts, theorizing that to describe a political candidate, for example, a person might generate a network-shaped picture regarding the candidate, which could be composed of various issues and attributes connecting to each other in his or her mind. With regard to the second assumption, an alternative view suggests that the news media can also transfer the salience of interrelationships between issues and attributes to the public’s mind. In her Limited Capacity Model, Lang16 discussed the process through which individuals link new messages received from the news media to their existing associative memory network. Once the connection is stored, whether the connection could be held firmly in the audience’s memory depends on whether the two constructs can be activated in tandem by internal or external factors, such as media coverage. Therefore, the more often news media report two issues during the same time period, the more recurrently two issues are activated jointly; thus, the greater chance the two’s interrelationship can be retrieved later. Drawing upon Lang’s17 model, the NAS model18 hypothesizes that how the news media associate issues and attributes influences the audience’s perception. In sum, moving beyond these two assumptions, the NAS model hypothesizes that the salience of interrelationships among objects and/or attributes can be transferred from the media agenda to the public agenda. As mentioned earlier, two empirical studies have been conducted to test the NAS model. Guo and McCombs19 analyzed the data initially collected for Kim and McCombs’s comprehensive analysis of attribute agenda-setting effects.20 Employing a social network analysis approach, the researchers compared the media and public agenda networks of the political candidate attributes in the Texas gubernatorial and U.S. senatorial elections in 2002. Results demonstrated statistically significant correlations between the two networks, and thus, the model was supported. Based on a new set of data, the second NAS study21 compared the media and public agenda networks regarding the political candidates’ qualifications and character in the 2010 Texas gubernatorial election. The findings again supported the NAS model. It should be noted that both studies focused on the network relationships among a list of attributes, whereas the NAS model could theoretically be applied to networks that consist of attributes, objects, or combinations of objects and attributes. In addition, given the fact that both exploratory studies were conducted in a local setting and focused on political elections, whether the model could also apply to the national level and other socio-cultural contexts remains a question. The present study seeks to address these two gaps by testing the model on object-based agenda networks and on a nationwide scale.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 672 Journalism & Mass Communication Quarterly 91(4)

Although the NAS and its network analysis statistical methodology are new aspects of agenda-setting theory that have emerged in the last two years, at its core, this is an extension of a gestalt perspective that has its roots in the earliest days of agenda setting. By a gestalt, we mean the collective mix of major public issues and news topics presented by the news media to the public. This gestalt perspective also describes what members of the public experience and absorb as they are exposed to the media agenda. The resulting gestalt for members of the public is described by Lippmann22 as “the pictures in our heads.” As individuals turn the pages of a newspaper, scroll, or click their way through an online site, or watch the news on television, they are presented with a jumble of information that is, at best, only partially organized. What makes this gestalt comprehensible to the public as they organize their picture of the world is the high degree of redundancy in the media agenda for major issues across media platforms and over time23 and the public’s exposure to multiple over-lapping platforms.24 And for agenda-setting effects, Stromback and Kiousis25 found that the total amount of attention to political news—attention to the gestalt presented by the media—predicted the salience of the issues most important to voters in deciding upon their vote, not their attention to specific news media. Whereas McCombs and Shaw26 compared recent news coverage with undecided voters’ issue agenda in Chapel Hill during the fall weeks of the 1969 presidential campaign, Funkhouser27 compared the major issues in the news with the public’s responses about the most important issues facing the country across an entire decade, a macro- complement to the micro–Chapel Hill study. The combination of these perspectives gave agenda setting a strong beginning. In terms of a gestalt perspective, the significant association (ρ = .78, p < .05) found by Funkhouser illustrates the robust fit between the collective symbolic representation of the world presented in the news— the media gestalt—and the collective picture of that world among the public—the public gestalt. As previously noted, traditional first- and second-level agenda-setting research examines the transfer of discrete message elements—issues or attributes— from the media to the public. However, regarding the media agenda as a bundle of networked elements—a gestalt—and examining the connections among these elements offer a more comprehensive view of what the media present and what the public experiences. Our initial goal here is to examine how strongly the media and public gestalts described in terms of networked interconnections among issues are associated with each other.

H1: The network relationships among issues on the media agenda will be positively associated with the public’s perceived importance of this object network.

Our second hypothesis assumes support for H1 and moves forward to examine the relationships among these media and public gestalts across time.

H2: Across time, the network relationships among issues on the media agenda influence the network relationships among issues among the public.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 673

Our third hypothesis sets up a comparison between the NAS model and traditional first-level agenda-setting analysis.

H3: At the first level of agenda setting, the public object agenda regarding the most important issues facing the country reflects the media object agenda.

Intermedia Network Agendas This study seeks to explore further the role of different media channels in mapping out the world. In addition to demonstrating a high degree of correspondence between the media and the public agendas, the Chapel Hill study28 also reported a high degree of correspondence (a median correlation of .71) among the agendas of the nine news media content analyzed. If the two news magazines are excluded, the median correlation of the seven newspaper and television agendas is .81. At the first level of agenda setting, this high degree of correspondence among media agendas has been exten- sively replicated over the years across the world.29 This study examines the degree of intermedia agreement regarding the issue agenda at the network level of analysis.

H4: Network media issue agendas are positively correlated among newspaper, online media, television, and radio.

Method Data Collection This study is based on a series of secondary data analyses. The media content analysis data were initially collected by the PEJ and the public survey data by the Gallup national polls. The hypotheses were tested on the five years from 2007 to 2011. A total of ten sub-datasets were used in this study: (1a) 2007 content analysis data, (1b) 2007 public opinion data, (2a) 2008 content analysis data, (2b) 2008 public opinion data, (3a) 2009 content analysis data, (3b) 2009 public opinion data, (4a) 2010 content analysis data, (4b) 2010 public opinion data, (5a) 2011 content analysis data, and (5b) 2011 public opinion data (see Table 1). Those five years included both nonelection and election years, which provides a broad view of the recent media and public agenda in the United States. Data availability was also a reason for the five-year selection. The media analysis data used for this study—the Pew PEJ database—only included analysis from 2007 to the present.

Content analysis. The Pew Research Center’s PEJ monitors weekly media content in the United States, detailing the most covered topics in four dozen news outlets, including newspapers, online news sites, network television, cable television, and radio.30 Many mass communication studies31 have used the PEJ news index to analyze media coverage. For this study, all the weekly reports were retrieved for 2007, 2008, 2009, 2010, and 2011. The unit of analysis was a PEJ weekly report of the news coverage.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 674 Journalism & Mass Communication Quarterly 91(4)

Table 1. Datasets for Analysis.

Year Content analysis Gallup public survey 2007 (1a) PEJ 49 weekly reports (1b) 12 monthly surveys 2008 (2a) PEJ 52 weekly reports (2b) 12 monthly surveys 2009 (3a) PEJ 52 weekly reports (3b) 12 monthly surveys 2010 (4a) PEJ 51 weekly reports (4b) 12 monthly surveys 2011 (5a) PEJ 49 weekly reports (5b) 12 monthly surveys

Note. PEJ = Project for Excellence in Journalism

Specifically, each PEJ weekly report listed the top ten topics ranked by their percentage of news coverage during the particular week. Coding of these weekly reports yielded a large number of specific topics that were then coded into ten major issue categories: (1) Economy, (2) Health, (3) Wars, (4) Politics, (5) National Security, (6) Social Disorder, (7) Education, (8) Environment, (9) Immigration, and (10) Other.32 The decision about these ten issue categories was based on the preliminary coding of the Gallup polls of the same years and previous agenda-setting studies.33 Two trained coders coded about 17% of the PEJ weekly reports to determine intercoder reliability. Tests showed satisfactory results with an achieved reliability coefficient of .98, Cohen’s kappa, a conservative measurement of reliability that takes into account the agreement occurring by chance.34 Variables where Cohen’s kappa is at or above .81 are considered “almost perfect agreement.”35 The coder who was not among the authors coded the rest of the data. The analysis for this study included the top five issue categories in each PEJ weekly report.36 If more than one topic listed in the weekly reports fell into the same issue category, the percentage of news coverage for each topic was summed to decide whether the issue category should be included in the analysis. For example, in one PEJ weekly report, “terrorism” received 15% of the news coverage, and “North Korea’s nuclear test” received 10%. Both of these two topics were coded into the category “National Security.” Adding up the percentage of the news coverage for both topics made the category “National Security” one of the top five issue categories of that week. In line with the gestalt perspective detailed above, the collective mix of news topics and major public issues that describes the information environment presented by the news media, the co-occurrence of top issues in each week’s PEJ report was the basic datum for the network analysis. Over the course of a year, some pairs of issues co- occurred frequently, others rarely or not at all. Details of the network analysis prepared from these data are detailed below.

Public survey. The public agenda data used in this study were based on the Gallup Poll results. Since 1939, Gallup has surveyed the general public, asking “What do you think are the most important problems facing the country today?”37 Numerous agenda- setting studies38 have examined public opinion by analyzing results from these Gallup polls. This study analyzed the Gallup polls from 2007 to 2011, with one poll every

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 675

month during the five years. As such, the public agenda data for this study consist of answers from a total of sixty surveys. The unit of analysis for the public survey data was a monthly poll. Using the same ten issue categories from the content analysis, the top five issue categories were coded for each monthly survey. The same two trained coders coded 20% of the public survey data and achieved an intercoder reliability of .93 with Cohen’s kappa. Again, consistent with the gestalt perspective regarding the pictures in the minds of the public, the co-occurrence of the top issues in a Gallup poll was the basic datum for the network analysis. Over the course of a year, some pairs of issues co-occurred frequently in the polls, others rarely or not at all. Details of the network analysis prepared from these data are given below.

Data Analysis To examine H1 and H2, which hypothesize a strong association between the media issue network agenda and the public issue network agenda and the agenda-setting influence of the media agenda on the public agenda, the study employs social network analysis as used in the previous NAS studies.39 The first step of network analysis was to generate matrices that respectively represent media agenda networks and public agenda networks for each year. For each one of the ten subsets of data, a symmetrical matrix of nine rows × nine columns was created. Each row or column represents an issue category. The category of “Other” was not included in the analysis. The entry in each cell is the frequency associated with the relationship between two issue categories. Specifically, to capture the news media gestalt, the collective mix of major public issues presented by the news media to the public each week, and the public agenda gestalt, the public’s collective view of the major public issues each month, the study measured the relationship between two issue categories by calculating their co-occurrence in the same PEJ weekly report or in the same monthly survey. The more frequently the two issue categories co-occurred across the PEJ’s weekly reports, the stronger their association on the media agenda. Similarly, taking the general public as a collective, the more often the two issue categories appeared across the same polls, the more connected the two issues were on the public’s mind, that is, in the public’s collective gestalt of the issues measured each month. Thus, the unit of analysis in this network analysis was a dyad: two issue categories and their relational ties. One matrix with a set of ties based on the PEJ content analyses from the weekly reports across the year defines the media network issue agenda. A separate matrix with a set of ties based on the monthly Gallup polls defines the public network issue agenda. Table 2, based on the 2007 content analysis data, and Table 3, based on the 2007 survey data, provide two examples of these matrices. In Table 2, for example, the entry in the cell corresponding to A (i.e., Economy) and B (i.e., Healthcare) is 7, which means the two issue categories “Economy” and “Healthcare” appeared together in 7 of the PEJ weekly reports in the year of 2007. Similarly, the entry in the cell

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 676 Journalism & Mass Communication Quarterly 91(4)

Table 2. Matrix of 2007 Media Object Agenda.

ABCDEFGHI A B 7 C 26 11 D 25 11 48 E 16 9 37 36 F 13 2 28 27 19 G 0 0 0 0 0 0 H 6 1 15 15 10 5 0 I 7 3 21 20 15 12 0 5

Note. A = Economy; B = Health; C = Wars; D = Politics; E = National Security; F = Social Disorder; G = Education; H = Environment; I = Immigration.

Table 3. Matrix of 2007 Public Object Agenda.

ABCDEFGHI A B 6 C 12 6 D 10 5 10 E 7 3 7 5 F 9 3 9 8 4 G 0 0 0 0 0 0 H 0 0 0 0 0 0 0 I 4 1 4 2 2 3 0 0

Note. A = Economy; B = Health; C = Wars; D = Politics; E = National Security; F = Social Disorder; G = Education; H = Environment; I = Immigration. corresponding to A (i.e., Economy) and B (i.e., Health) in the public agenda matrix is 6 (see Table 3), indicating that these two issues were mentioned together in 6 of the monthly Gallup polls in 2007. The researchers then compared the media agenda networks and public agenda networks to explore whether the two agendas were correlated. Specifically, Quadratic Assignment Procedure (QAP) correlation and regression tests were conducted to compare the five pairs of matrices: (1a-b) 2007 content analysis and survey, (2a-b) 2008 content analysis and survey, (3a-b) 2009 content analysis and survey, (4a-b) 2010 content analysis and survey, and (5a-b) 2011 content analysis and survey. The QAP correlation test computes the correlation between entries of two square matrices. The QAP regression “test” is to regress a dependent matrix on one or more independent matrices, and assess the significance of the r-square and regression coefficient.40 In other words, the use of the QAP regression test can further examine whether one

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 677

matrix—the independent matrix—can predict another—the dependent matrix.41 In the present regression analyses, content analysis results were the independent matrices and survey results were dependent matrices. The present study also employed cross-lagged correlation analysis, which was first tested by Campbell to identify the directional causality between the media and the public agendas.42 Instead of testing the relationship between variable X and variable Y in the same time period as in traditional correlation, cross-lagged analysis examines these two variables at two different points of time. Specifically, variable X at Time 1 (Xt1) is correlated with variable Y at Time 2 (Yt2), and variable X at Time 2 (Xt2) is correlated with variable Y at Time 1 (Yt1). The correlation coefficients are then compared with one another as well as with the Rozelle-Campbell baseline value, r(Xt1, Yt2) versus r(Xt2, Yt1). If r(Xt1, Yt2) is larger than r(Xt2, Yt1), and r(Xt1, Yt2) is above while r(Xt2, Yt1) is below the baseline, it can be reasoned that X has stronger effects on Y than the reverse. Cross-lagged correlation has been used widely in previous agenda-setting research in similar fashion.43 This study divided the data into six-month periods for both the media and the public agendas. Across the entire five-year period, 2007 to 2011, there were a total of ten periods of time to be tested with nine cross-lagged correlations (see Figure 3). For example, we correlated the media agenda data in the first half of 2007 with the public agenda data in the second half of 2007, and the media agenda data of the second half with the public agenda data in the first half of that year. To test H3, which compares the third-level and the first-level agenda-setting effects, traditional rank-order correlations of the media object agenda and public object agenda for each year were calculated (see Table 4). To examine H4 that explores the associations among network media agendas of different media outlets, another set of matrices were created to represent media network issue agendas of newspapers, cable television, network television, online news sites, and radio for each year. QAP correlation tests were conducted to explore simi- larities among different media outlets’ matrices.

Results For H1, which hypothesizes positive associations between the media and public network issue agendas, the results demonstrated statistically significant associations between the public network agendas and the media network agendas in all five years (see Table 5). Specifically, QAP correlation and regression results showed that the coefficients were the highest in 2009 (Pearson’s r = .87, p < .01, R2 = .76, p < .01), and the lowest in 2011 (Pearson’s r = .65, p < .05, R2 = .42, p < .05). H1 was supported. To illustrate these issue networks visually, Figures 1 and 2 present the media and public networks in 2007, as an example. Lines show the connections between issues. The number of connections an issue has with others determines how central it is in the graphs. The strength of the association between two issues is illustrated by the thick- ness of the line. For example, the two graphs indicate that Wars, Economy, and Politics are the three most central issues in both the media and the public agendas in 2007.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 678 Journalism & Mass Communication Quarterly 91(4)

Table 4. Frequency of Issues on the Media and Public Object Agendas.

2007 2007 2008 2008 2009 2009 2010 2010 2011 2011 Objects media public media public media public media public media public Economy 26 12 49 12 52 12 51 12 49 12 Health 11 6 5 5 36 12 26 11 11 11 Wars 49 12 38 12 42 10 20 2 14 0 Politics 48 10 52 12 43 12 49 12 46 12 National 37 7 38 9 41 7 28 11 43 12 security Social 28 9 26 9 12 7 20 7 25 12 disorder Education 0 0 0 0 3 0 9 1 19 1 Environment 16 0 23 0 11 0 27 2 16 0 Immigration 21 4 5 1 1 0 13 2 5 0 Other 11 0 13 24 21 0 13 0 16 0

Table 5. Statistical Results Comparing Media and Public Agendas.

Correlation QAP correlation QAP Dataset (Spearman’s ρ) (Pearson’s r) regression (R2) 2007 .79* .75** .56** 2008 .88** .84** .71** 2009 .87** .87** .76** 2010 .87** .72** .52** 2011 .82** .65* .42* 5-Year .81** .81** .65**

Note. QAP = Quadratic Assignment Procedure. *p < .05. **p < .01.

Education in both graphs, however, is at the margin, having no connection with other issues. To determine the direction of causality between the two agendas, cross-lagged correlations were calculated (see Figure 3). Results indicated that of the nine cross-lagged correlations, four (2007-2008, 2008 [1st half vs. 2nd half], 2010 [1st half vs. 2nd half], and 2010-2011) show clear influence of the media agenda on the public agenda. For example, the coefficient of the correlation between the media agenda in the second half of 2007 and the public agenda in the first half of 2008 (r =.84, p < .01) was stronger than that between the media agenda in the first half of 2008 and the public agenda in the second half of 2007 (r = .49, p > .05). The Rozelle-Campbell baseline was .54. Two other cross-lagged analyses (2007 and 2009-2010) showed a stronger influence of the media agenda on the public agenda than the reverse, but both correlations were above the baseline. Only two cross-lagged analyses (2008-2009 and

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 679

Figure 1. Media object agenda network, 2007.

2009) indicated stronger effects of the public agenda on the media agenda than the reverse, but all correlations were above the baseline. One year saw a reciprocal influence (2011). In short, two-thirds of the cross-lagged analyses supported H2 that across time, the network relationships among issues on the media agenda influenced the network relationships of these issues among the public. Taken together, the results for H1 and H2 support the NAS model that the salience of network relationships among issues in the news media can be transferred to the public’s mind. For H3, which sets up a comparison between traditional first-level agenda-setting effects and these third-level effects, the rank order of the media issue agenda significantly correlated with the public issue agenda in all five years. The highest Spearman’s rho coefficient was in 2008 (ρ = .88, p < .01, see Table 5), whereas the lowest was in 2007 (ρ = .79, p < .05). These first-level measurements are similar to the third-level network measures. The largest difference is for 2011 (.82 measured at the first level, .65 measured at the network level), but both show a strong association between the media and public issue agendas.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 680 Journalism & Mass Communication Quarterly 91(4)

Figure 2. Public object agenda network, 2007.

Interestingly, when the data for all five years are aggregated, the correlations are identical for the two measures. Both are .81. And these two measures comparing the media and public issue gestalts from 2007 to 2011 are almost identical to Funkhouser’s44 comparison of these issue gestalts from the 1960s (ρ = .78, p < .05). For H4, which considers the associations among media issue agendas of different media channels, the results demonstrated statistically significant correlations among all the media agenda networks constructed by newspapers, television news, online media, and radio (see Table 6). Specifically, in 2007, the QAP correlation coefficients ranged from r = .82, p < .01 (network/cable television) to r = .96, p < .001 (newspaper/ radio, and network television/radio). In 2008 the lowest was the correlation between cable television and newspaper, with the coefficient at r = .73, p < .01, whereas the highest was between radio and online: r = .99, p < .001. In 2009, the correlation coefficients ranged between r = .87, p < .001 (newspaper vs. network television) and r = .97, p = .001 (newspaper vs. Online, online vs. cable television, and cable television vs. radio). In 2010, the lowest correlation coefficient was r = .72, p < .01 (newspaper versus cable television). The highest was r = .96, p < .001. In 2011, newspaper was

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 681

Media Media Media Media Media Media 1st half 2007 2nd half 2007 2nd half 2007 1st half 2008 1st half 2008 2nd half 2008 .89** .77** 1***

.82** .84** .98***

.63** .60* .60* .98*** .98*** .80**

.55* .49 .80**

.79** .59* .87** Gallup Gallup Gallup Gallup Gallup Gallup 1st half 2007 2nd half 2007 2 nd half 2007 1st half 2008 1 st half 2008 2nd half 2008

Rozelle-Campbell baseline= .43 Rozelle-Campbell baseline= .54 Rozelle-Campbell baseline= .80

Media Media Media Media MediaMedia 2 2nd half 2008 1st half 2009 1st half 2009 2nd half 2009 2nd half 2009 1st half 2010 .89** . 71*

.62* .54* . 75**

.67* .80** .73** .73** .85**

.85** .78** .80** .93**

.56* .76* .96** Gallup Gallup Gallup Gallup nd st 2 half 2008 1 half 2009 1st half 2010 2nd half 2010 .67* Gallup Gallup nd st Rozelle-Campbell baseline= .52 Rozelle-Campbell baseline= .44 2 half 2009 1 half 2010

Rozelle-Campbell baseline= .50

Media Media Media Media Media Media 1st half 2010 2 nd half 2010 2nd half 2010 1st half 2011 1st half 2011 2 nd half 2011 . 75** . 71** .88**

. 75** . 69* .61*

.78** .57* .57* .69* .69* .58*

.54 .48 .64*

.95* * .93* * .99** Gallup Gallup Gallup Gallup Gallup Gallup 1st half 2010 2nd half 2010 2nd half 2010 1st half 2011 1st half 2011 2nd half 2011

Rozelle -Campbell baseline= .44 Rozelle-Campbell baseline= .40 Rozelle-Campbell baseline= .40

Figure 3. Cross-lagged correlation of the media and the public agendas. *p < .05. **p < .01. ***p < .001. least correlated with network television, with the coefficient being at r = .71, p < .05. Radio and cable television saw the strongest association, r = .98, p < .001. H4 was strongly supported.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 682 Journalism & Mass Communication Quarterly 91(4)

Table 6. Correlations among Network Media Agendas of Different Media Outlets from 2007 to 2011.

Media sector 2007 2008 2009 2010 2011 Newspaper vs. online .86** .90*** .97*** .73** .76* Newspaper vs. network .91*** .83*** .87*** .82** .71* Newspaper vs. cable .88*** .73** .95*** .72** .78** Newspaper vs. radio .96*** .92*** .92** .74** .86** Online vs. network .96*** .97*** .92*** .92*** .96*** Online vs. cable .87** .82** .97*** .94*** .98*** Online vs. radio .94*** .99*** .93*** .96*** .95*** Network vs. cable .82** .79** .94*** .83** .94*** Network vs. radio .96*** .96*** .95*** .91*** .91*** Cable vs. radio .90*** .85** .97*** .93*** .97***

*p < .05. **p < .01. ***p < .001.

Discussion and Conclusion Comparing media issue agenda networks based on the PEJ weekly news reports and public issue agenda networks based on the Gallup Polls in five years from 2007 to 2011, this study found significant associations between these agendas in all five years, as well as strong evidence of significant agenda-setting effects in the cross-lagged analyses, supporting a core proposition of third-level agenda setting that the salience of the network relationships among issues can be transferred from the news media to the public. Notably, this study contributed theoretically to the NAS by expanding the scope of research. Unlike the two initial NAS studies,45 which were conducted in a local setting during election periods, the current study investigated third-level agenda setting at a national scale during much longer periods of time. In addition, while the early stud- ies46 tested the NAS model on a set of attributes—political candidates’ character and qualifications—this study applied the model to the object agenda. As such, this study provided not only an entirely new set of empirical evidence to support the NAS model, but also advanced the model’s theoretical scope. Equally important, this study demonstrated that the third-level network effects parallel the first level of agenda-setting effects in all five years. Moreover, this study found that the media issue agenda networks were extremely similar across various media outlets including newspaper, radio, television, and online news media. In other words, different news media channels are not only similar in terms of the traditional hierarchical agenda, but they also are highly similar in terms of network issue agenda. In the earliest days of agenda-setting research, the Chapel Hill study47 found a high degree of homogeneity among media issue agendas. This study extends that finding to the third level of agenda setting. In particular, it is interesting to note that the media network agendas constructed by online news media had the strongest correlations with those constructed by other media outlets. This finding may

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 683

reflect the medium’s increasingly dominant role in today’s media landscape, through providing unlimited space where all traditional mediums intersect. To conclude, this study improved the NAS in various aspects, solidly illustrating the predictive power of the third-level agenda-setting effect. This effect is a much broader and more comprehensive perspective about media effects than traditional agenda-setting hypotheses focused on the salience of discrete objects and attributes. This focus at the third level on bundles of message elements provides a more nuanced understanding of agenda-setting effects and further enriches our understanding of Lippmann’s48 metaphor that the news media construct a pseudo-environment, a gestalt, for the public bridging “the world outside and the pictures in our heads.”49 The news is experienced in bits and pieces, but over time coalesces into a gestalt that strongly influences the collective public gestalt, the pictures in our minds. Network analysis helps us capture the nuances of those gestalts. Issues do not exist in isolation in the news. They co-exist in the daily information environment with other issues, and network analysis captures this larger picture of each issue in the context of that environment. In particular, network analysis’ measure of centrality describes how often each issue appears over time with other issues on the media agenda. Still, this study has several limitations. First, by examining how frequently the public mentioned two issues together, the study could only identify how the public associated the issues implicitly. Future studies should employ the mind-mapping survey method,50 asking respondents to draw connections between issues and attributes by themselves. In other words, in addition to the macro-analyses here, which extend Funkhouser’s51 and later studies to the network level, future research should also examine these networks at the micro-level. Second, the study tested the NAS model based on aggregated secondary data. More studies based on original data should be conducted to explore object-based media and public agenda networks. Finally, the use of a small number of issues in this study might have created a potential spurious relationship between the two agendas. Future research should investigate this relationship with larger sets of topics.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. Lei Guo and Maxwell McCombs, “Network Agenda Setting: A Third Level of Media Effects” (paper presented at the annual conference of the International Communication Association, Boston, MA, May 2011); Lei Guo and Maxwell McCombs “Toward the Third-Level Agenda-Setting Theory: A Network Agenda-Setting Model” (paper presented at the annual conference of the AEJMC, St. Louis, MO, 2011); Lei Guo, “Toward the Third

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 684 Journalism & Mass Communication Quarterly 91(4)

Level of Agenda-Setting Theory: A Network Agenda-Setting Model” in Agenda Setting in a 2.0 World: A Tribute to Maxwell McCombs, ed. Thomas Johnson (NY: Routledge, 2013), 112-33. 2. Guo, “Toward the Third Level,” 2. 3. Walter Lippmann, Public Opinion (NY: Macmillan, 1922). 4. Lei Guo, “The Application of Social Network Analysis in Agenda Setting Research: A Methodological Exploration,” Journal of Broadcasting & Electronic Media 56 (4, 2012): 616-31. 5. Guo and McCombs, “Network Agenda Setting”; Guo and McCombs “Toward the Third- Level Agenda-Setting.” 6. Maxwell McCombs and Donald L. Shaw, “Agenda-Setting Function of Mass Media,” Public Opinion Quarterly 36 (2, 1972): 176-87. 7. Maxwell McCombs, Setting the Agenda: The Mass Media and Public Opinion (Cambridge, MA: Polity Press, 2004). 8. Maxwell McCombs, Esteban Lopez-Escobar, and Juan P. Llamas, “Setting the Agenda of Attributes in the 1996 Spanish General Election,” Journal of Communication 50 (2, 2000): 77-92; Donald L. Shaw and Maxwell McCombs, The Emergence of American Political Issues: The Agenda-Setting Function of the Press (St. Paul, MN: West, 1977); David Weaver, Doris A. Graber, Maxwell E. McCombs, and Chaim H. Eyal, Media Agenda Setting in a Presidential Election: Issues, Images, and Interest (NY: Praeger, 1981). 9. Wayne Wanta, Guy Golan, and Cheolhan Lee, “Agenda Setting and International News: Media Influence on Public Perceptions of Foreign Nations,” Journalism & Mass Communication Quarterly 81 (2, 2004): 364-77. 10. Guo and McCombs, “Network Agenda Setting”; Guo and McCombs, “Toward the Third- Level Agenda-Setting.” 11. Stephen Kaplan, “Cognitive Maps in Perception and Thought,” in Image and Environment: Cognitive Mapping and Spatial Behavior, ed. Roger M. Downs and David Stea (Chicago: Aldine, 2005), 63-78. 12. John R. Anderson and Gordon H. Bower, Human Associative Memory (Washington, DC: Winston, 1973); John R. Anderson, The Architecture of Cognition (Cambridge, MA: Harvard University Press, 1983). 13. Kaplan, “Cognitive Maps in Perception,” 2. 14. Eric L. Santanen, Robert O. Briggs, and Gert-Jan de Vreede, “The Cognitive Network Model of Creativity: A New Causal Model of Creativity and a New Brainstorming Technique” (paper presented at the 33rd Hawaii International Conference on System Sciences, Los Alamitos, CA, January 2000). 15. Guo and McCombs, “Toward the Third-Level Agenda-Setting.” 16. Annie Lang, “The Limited Capacity Model of Mediated Message Processing,” Journal of Communication 50 (1, 2000): 46-71. 17. Lang, “The Limited Capacity Model.” 18. Guo and McCombs, “Toward the Third-Level Agenda-Setting”; Guo, “The Application of Social Network Analysis.” 19. Guo and McCombs, “Network Agenda Setting.” 20. Kihan Kim and Maxwell McCombs, “News Story Descriptions and the Public’s Opinions of Political Candidates,” Journalism & Mass Communication Quarterly 84 (2, 2007): 299-314. 21. Guo and McCombs, “Toward the Third-Level Agenda-Setting.” 22. Lippmann, Public Opinion.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Vu et al. 685

23. McCombs and Shaw, “Agenda-Setting Function”; Pablo J. Boczkowski, News at Work: Imitation in an Age of Information Abundance (Chicago: University of Chicago Press, 2010); Maxwell McCombs and Jian H. Zhu, “Capacity, Diversity and Volatility of the Public Agenda: Trends from 1954 to 1994,” Public Opinion Quarterly 59 (4, 1995): 495-525. 24. James G. Webster and Thomas B. Ksiazek, “The Dynamics of Audience Fragmentation: Public Attention in an Age of Digital Media,” Journal of Communication 62 (1, 2012): 39-56. 25. Jesper Stromback and Spiro Kiousis, “A New Look at Agenda Setting Effects— Comparing the Predictive Power of Overall Political News Consumption and Specific News Media Consumption across Different Media Channels and Media Types,” Journal of Communication 60 (2, 2010): 271-92. 26. McCombs and Shaw, “Agenda-Setting Function.” 27. Ray G. Funkhouser, “The Issues of the Sixties,” Public Opinion Quarterly 37 (1, 1973): 62-75. 28. McCombs and Shaw, “Agenda-Setting Function.” 29. Marilyn Roberts and Maxwell McCombs, “Agenda Setting and Political Advertising: Origins of the News Agenda,” Political Communication 11 (3, 1994): 249-62; Craig Trumbo, “Longitudinal Modeling of Public Issues: An Application of the Agenda-Setting Process to the Issue of Global Warming,” Journalism Monographs 152 (August 1995): 1-57; Esteban Lopez-Escobar, Juan P. Llamas, Maxwell McCombs, and Federico R. Lennon, “Two Levels of Agenda Setting among Advertising and News in the 1995 Spanish Elections,” Political Communication 15 (2, 1998): 225-38; Boczkowski, News at Work. 30. “Examining The Changing News Landscape,” Pew Research Journalism Project, http:// www.journalism.org/about/ (accessed August, 2014). 31. Jong H. Lee, “News Values, Media Coverage, and Audience Attention: An Analysis of Direct and Mediated Causal Relationships,” Journalism & Mass Communication Quarterly 86 (1, 2009): 175-90; Scott Maier, “All the News Fit to Post? Comparing News Content on the Web to Newspapers, Television, and Radio,” Journalism & Mass Communication Quarterly 87 (3-4, 2010): 548-62. 32. (1) “Economy” includes any economic or financial issues such as inflation, economic crisis; (2) “Health” includes such issues as healthcare debate or general discussions on health; (3) “Wars” consists of media coverage on the Iraq War and the war in Afghanistan; (4) “Politics” refers to coverage of the government and congress; (5) “National Security” comprises issues on terrorism, national security, and international conflicts; (6) “Social Disorder” includes issues related to crimes, moral degradation, and other social problems; (7) “Education” refers to education or any related issues; (8) “Environment” houses the coverage about general environmental issues or natural disasters; and (9) “Immigration” includes issues related to immigration in the United States. 33. Stephanie Craft and Wayne Wanta, “Women in the Newsroom: Influences of Female Editors and Reporters on the News Agenda,” Journalism & Mass Communication Quarterly 81 (1, 2004): 124-38; Spiro Kiousis, “Explicating Media Salience: A Factor Analysis of the New York Times Issue Coverage during the 2000 U.S. Presidential Election,” Journal of Communication 54 (1, 2004): 71-87. 34. Jacob Cohen, “A Coefficient of Agreement for Nominal Scales,” Educational and Psychological Measurement 20 (1, 1960): 37-46. 35. Anthony J. Viera and Joanne M. Garrett, “Understanding Interobserver Agreement: The Kappa Statistic,” Family Medicine 37 (5, 2005): 360-63.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 686 Journalism & Mass Communication Quarterly 91(4)

36. Although the Project for Excellence in Journalism (PEJ) weekly reports list ten topics, many of these are specific to the moment or are fleeting mentions of other issues. To obtain a picture of the broad sweep of media coverage and public opinion, the five master issues listed here are used. This is similar to the strategy in McCombs and Zhu’s examination of the Gallup MIP or Most Important Problems question across a forty-year period. See McCombs and Zhu, “Capacity, Diversity, and Volatility.” 37. David S. Yeager, Larson B. Samuel, Jon A. Krosnick, and Trevor Tompson, “Measuring Americans’ Issue Priorities: A New Version of the Most Important Problem Question Reveals More Concern about Global Warming and the Environment,” Public Opinion Quarterly 75 (1, 2011): 125-38; Tom W. Smith, “America’s Most Important Problem—a Trend Analysis, 1946-1976,” Public Opinion Quarterly 44 (2, 1980): 164-80. 38. Esteban Lopez-Escobar, Juan P. Llamas, and Maxwell McCombs, “Agenda Setting and Community Consensus: First- and Second-Level Effects,” International Journal of Public Opinion Research 10 (4, 1998): 335-48; Maxwell McCombs, “Building Consensus: The News Media’s Agenda-Setting Roles,” Political Communication 14 (4, 1997): 433- 43; Wayne Wanta and Yu-Wei Hu, “The Agenda-Setting Effects of International News Coverage: An Examination of Differing News Frames,” International Journal of Public Opinion Research 5 (3, 1993): 250-64. 39. Guo, “The Application of Social Network Analysis.” 40. Stephen P. Borgatti, Martin G. Everett, and Lin C. Freeman, Ucinet for Windows: Software for Social Network Analysis (Boston, MA: Analytic Technologies, 2002). 41. Robert A. Hanneman and Mark Riddle, eds., Introduction to Social Network Methods (Riverside: University of California, Riverside, 2005), http://www.faculty.ucr. edu/%7Ehanneman/nettext/ (accessed July 15, 2013). 42. Donald T. Campbell, “From Description to Experimentation: Interpreting Trends as Quasi-Experiments,” in Problem in Measuring Change, ed. Chester W. Harris (Madison: University of Wisconsin Press, 1963), 212-42. 43. Ying R. Du, “Intermedia Agenda-Setting in the Age of Globalization: A Multinational Agenda-Setting Test,” Global Media and Communication 9 (1, 2012): 19-36; Jeongsub Lim, “A Cross-Lagged Analysis of Agenda Setting among Online News Media,” Journalism & Mass Communication Quarterly 83 (2, 2006): 298-312; Lopez-Escobar et al., “Two Levels of Agenda Setting,” 225-38; Roberts and McCombs, “Agenda Setting and Political Advertising.” 44. Funkhouser, “The Issues of the Sixties.” 45. Guo and McCombs, “Network Agenda Setting”; Guo and McCombs, “Toward the Third- Level Agenda-Setting.” 46. Guo and McCombs, “Network Agenda Setting”; Guo and McCombs, “Toward the Third- Level Agenda-Setting.” 47. McCombs and Shaw, “Agenda-Setting Function.” 48. Lippmann, Public Opinion. 49. Lippmann, Public Opinion. 50. Guo and McCombs, “Toward the Third-Level Agenda-Setting”; Guo, “The Application of Social Network Analysis.” 51. Funkhouser, “The Issues of the Sixties.”

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 JMQXXX10.1177/1077699014550095Journalism & Mass Communication QuarterlyHyun and Moon 550095research-article2014

Audience Motivations and Responses

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 687–705 News Media’s Role in the © 2014 AEJMC Reprints and permissions: Issue-Voting Process: News sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550095 Attention, Issue Proximity, jmcq.sagepub.com and Vote Choice

Ki Deuk Hyun1 and Soo Jung Moon2

Abstract Synthesizing research documenting the effects of news use on learning and evaluation of policy issues, this study explores the role of news media in the issue-voting process, which involves cognitive, attitudinal, and behavioral elements. Data analysis from the 2008 American National Election Study (ANES) shows that news attention predicts issue importance, issue knowledge, and perception of closeness on issue positions for one presidential candidate over another, which relates to vote choice. Results also indicate that the role of news attention for issue voting tends to be stronger among political independents highly attentive to news media than for others.

Keywords issue proximity, issue voting, news, presidential election

A standard form of voting in a democracy requires voters to align their policy preferences with competing candidates’ policy platforms. Such ideal policy-conscious voting, called issue voting, motivates candidates to respond to public concerns and encourages elected officials to fulfill their campaign promises. Because the idea of issue voting presumes a rational electorate and responsive government, it is praised as superior to voting decisions based on candidate images or usual party affiliation.1 Studies have identified key individual characteristics relating to issue-voting behavior. Compared with those who are less affected by issue considerations, issue voters tend to be better educated, as well as more politically informed and involved.2

1Grand Valley State University, Allendale, MI, USA 2University of West Georgia, Carrollton, GA, USA Corresponding Author: Ki Deuk Hyun, Grand Valley State University, 290 Lake Superior Hall, Allendale, MI 49401, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Downloaded from http://www.elearnica.ir 688 Journalism & Mass Communication Quarterly 91(4)

While voters’ political resources, or lack thereof, have drawn considerable scholarly attention in the issue-voting literature, the role of news media in this type of voting process has been relatively neglected. Considering that the idea of issue voting is pre- mised on an informed electorate, news media as a main source of campaign information deserve scholarly attention. In addition, recent documentation of growing polarization in U.S. politics adds another important dimension to the discussion of the media’s role in issue voting. One strand of thought in support of rising issue voting has been anchored on the assumption of an increase of a politically independent and better-educated public in Western democracies who may bypass party cues and rely instead on news media to make autonomous voting decisions based on issue-relevant information.3 In the United States, the number of political independents has continued to increase over several decades, reaching its highest level in recent years. In a 2012 survey, about 38% of respondents reported themselves to be independents, almost doubling from 19% in 1940 and showing a noticeable growth from 30% in 2000.4 However, a growing body of research suggests a countertrend: that partisan tenden- cies, in the form of polarization, have risen among the U.S. electorate over the last few decades.5 Diversity of information sources along with growth of selective exposure6 further complicates the relationship between news media and audience political orientations in issue voting. These changes in the political and media environments require us to re-examine how voters’ partisanship and media use interact to affect issue voting. Previous studies provide scattered evidence suggesting the potential influence of news use on issue voting. News use has been found to contribute to public perception of issue importance7 and issue knowledge.8 Media emphasis on issues has also been associated with candidate evaluation and voting behaviors.9 Theoretical efforts to connect these findings under the overall framework of issue voting, however, have rarely been made. To address this oversight, we examine the relationship between individual news use and the issue-voting process. Based on the traditional view of issue voting as a multiplicative series of actions,10 we argue that news media have an important function in this process—increasing voters’ issue-importance perception, enhancing issue knowledge, and aiding voters’ calculation of congruence between their own policy preferences and competing candidates’ policy stands (issue proximity)—all of which should consequently translate into voting choices. We also examine whether news use has a stronger association with issue voting among political independents than among partisans, consistent with the image of issue voting as an autonomous electoral choice that relies on information obtained mainly from news media. For an empirical test, we analyzed data from the 2008 American National Election Study (ANES).

Literature Review Conditions of Issue Voting Although issue voting is touted as an ideal voting type in a representative democracy, early voting studies reported that only a minority of the public satisfied the criteria

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 689

necessary for issue voting. Most voters lacked basic preconditions such as issue interest and knowledge, stable policy preferences, and ability to link their policy orientations to their votes.11 While some later studies confirmed this unflattering image of the electorate,12 others discovered the impact of issues on electoral choices and outcomes.13 Aside from debates about the extent of issue voting, the literature commonly implies that at the individual level, it is information-oriented and motivated voting behavior. Issue voting presumes voters who are informed and concerned about policy issues, qualities that should be used as the main basis for candidate choice. Researchers have examined the role of news media in issue voting from two different approaches: the agenda-setting tradition and a more generalist perspective documenting the influence of news media on the elements of issue voting. Agenda-setting research has shown that voters perceive issues emphasized in the news media to be more important,14 and prioritize those issues for candidate evaluations15 and voting decisions.16 On the other hand, the generalist approach has focused on the overall contribution of news use, regardless of issue salience, to issue knowledge,17 issue-importance perception,18 and issue voting.19 The current research synthesizes the two theoretical approaches to the role of news media in the process. Following the common conceptualization from issue-voting literature, we regard issue voting as involving the consecutive process of attainment of issue knowledge and perception of issue importance. Then, issue knowledge and importance should guide voters’ evaluations about the agreement between their own and candidates’ issue positions, which ultimately influences vote choice. Research on the consequences of agenda setting20 provides a theoretical framework delineating the potential role of news media in the issue-voting process: news attention consecutively leads to cognitive (issue knowledge and issue importance), attitudinal (issue proximity), and behavioral consequences (vote choice). Unlike agenda setting, which focuses on the ordering of issue salience and its subsequent effects, however, we take a generalist approach that addresses overall contribution of news use to the fulfillment of basic cognitive and attitudinal preconditions leading to candidate choice.

News Use and Issue Importance The issue-voting process presumes that a voter cares about policy issues and has the motivation to use policy alternatives as standards for an electoral choice. Issue importance, as an amount of concern given to policy issues, can be considered as an essential attitudinal element in initiating issue voting. Importance attitude, in general, has been documented as guiding information processing regarding the attitude object and affecting subsequent attitudes and behaviors. Importance perception inspires people to gather and use attitude-relevant information to make decisions and take actions.21 Due to this activating function of importance, it is often emphasized as a motivator in the attitude-behavior connection. When people assign importance to an attitude object, they are “motivated to protect it, express it, and be faithful to it in action.”22

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 690 Journalism & Mass Communication Quarterly 91(4)

Sources of importance attitudes can be both individual and social, as well as long- standing and temporal. Although personal values, interests, and group identification have enduring influence on people’s importance attitudes, a drastic increase in issue salience, caused by an influx of new information, can alter the extent of citizen concerns about specific issues.23 News media, as a main channel of electoral information, can heighten the level of issue importance in a relatively short time period. Exposure to news not only structures the ordering of importance given to different issues,24 but also simultaneously increases the overall salience of multiple issues.25 Drawing on past research linking news use and issue importance, our first hypothesis posits the following:

H1: News attention will be positively related to issue importance.

News Use and Issue Knowledge Aside from voters’ issue concerns, the issue-voting process requires individual cognitive skill to recognize and compare different candidates’ issue positions. During electoral campaigns, the news media are expected to provide substantive information regarding candidates’ policy positions.26 Although studies have found the relationship between news use and acquisition of issue knowledge to be somewhat mixed, an individual’s news use tended to be related to the level of issue knowledge when it was measured on the left-right scale in which survey respondents were asked to place candidates’ relative issue positions, rather than when they were directly asked to answer candidates’ issue stands.27 This observation implies that news use is likely to contribute to voter approximation of candidate stances on various policy issues, even if voters may have difficulty articulating those exact stands. At the very least, the electorate may learn from the news media a minimum level of issue knowledge necessary for issue voting, enabling voters to place candidates’ issue positions correctly in relative left-right terms. Therefore, we predict that news use should contribute to voter issue knowledge.

H2: News attention will be positively related to issue knowledge.

News Use and Issue Voting The next condition for issue voting is that the electorate should be able to assess which candidates’ issue positions more closely match their own. The spatial theory of voting demonstrates how voter–candidate agreement about issue positions affects issue-voting behavior. Because voters perceive greater utility from a candidate whose policy positions are nearest to their own, they are more likely to vote for that candidate.28 In the issue-voting literature, the concept of issue proximity taps into the extent of voter– candidate agreement on policy stands. Issue proximity captures whether voters are on the same side of an issue position and to what extent they agree or disagree with different candidates’ policy stands.29

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 691

Voter assessment of issue proximity should involve voters’ motivation and skills in placing their own issue orientations against competing candidates’ stances for comparison. The importance an individual ascribes to policy issues can serve as the main motivator to engage in this cognitive effort. Issue importance inspires people to seek out and elaborate on issue-relevant information necessary for issue-based electoral choices. People tend to use important issues as criteria for candidate evaluation because these issues are cognitively more accessible.30 Voters are also more likely to detect greater differences in candidates’ issue positions when they perceive certain issues to be important. Issue importance, therefore, can either mediate or moderate issue voting.31 Because issue importance helps voters discern differences in candidates’ policy positions, we expect that it should have a positive relationship with issue proximity that reflects voters’ evaluation of the degree of agreement and disagreement between their policy preferences and different candidates’ policy alternatives. Therefore, our next hypothesis is as follows:

H3: Issue importance will be positively related to issue proximity.

Even if voters obtain issue-relevant information, they still should be able to process that information to translate it into issue voting. When voters lack knowledge about candidates’ stands, they are unable to assess rationally which candidate’s policy platforms correspond more closely with their own. Previous studies have observed that cognitive ability, such as level of political knowledge and expertise, is one of the preconditions for issue voting. Researchers have shown that those informed about politics in general,32 and knowledgeable about candidates’ specific policy positions,33 tend to rely more on issues to determine how to vote. The relationship between issue knowledge and issue proximity is a crucial test to validate the premise that issue voting is driven by individual evaluation of policy options. Perception of close issue proximity itself can happen without any consideration of candidates’ objective policy positions. Voters might decide that one candidate’s issue positions more closely correspond to their own simply because they prefer that candidate over other candidates. If issue voting is a rational decision made on the basis of information, as normative theory assumes, it should be correlated with issue knowledge. Therefore, our next hypothesis is as follows:

H4: Issue knowledge will be positively related to issue proximity.

Although previous research has explored the impact of news use on issue importance and issue knowledge, few studies have extended their findings to voter assessment of the match between their policy preferences and candidate positions. It is necessary to include that weighing process of voter–candidate issue agreement, because it constitutes a key aspect of meaningful issue voting. We suppose that news media, as a major source of campaign-related information, can exert a significant influence on people’s evaluations of issue proximity.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 692 Journalism & Mass Communication Quarterly 91(4)

As a concept linking voter–candidate issue positions, issue proximity entails the intensity or strength of voters’ opinions regarding the degree of closeness between their own preferences and the positions of competing candidates.34 Because attention to information tends to strengthen attitudes and evaluations,35 we predict that news attention would lead to greater intensity of issue proximity. Voters will feel greater closeness to one candidate’s issue positions, while perceiving greater distance to the other candidate’s positions from their own preferences.

H5: News attention will be positively related to issue proximity.

Finally, we raised one research question regarding the interaction of news use and partisanship on issue voting. Because issue voting assumes independent and rational voting decisions by voters, rather than habitual choice based on party or group affinity, it is often highly regarded. In this respect, declining partisanship in Western democracies has raised the expectation of an upsurge in issue voting. Because a significant portion of the public disassociates itself from party loyalty, the room for issue impact on voting choice increases accordingly. Also, greater availability and diversity of news media in contemporary campaigns facilitate issue voting by providing information necessary for voting choice, based on the merits of policy alternatives instead of partisan affiliation.36 Although recent studies provide evidence of greater partisan polarization, at least in U.S. politics,37 the basic thesis may hold true that an issue voter is politically unaffiliated and media-reliant. If issue voting indeed represents an individual voter’s autonomous, information-oriented choice, news media should have greater impact on political independents who are unconstrained by party affinity.

RQ1: Does issue proximity have a stronger relationship with vote choice among political independents who are more attentive to news than others?

Method Data The 2008 ANES was employed for data analysis. All questions in the analysis, except attention to the Internet for campaign information and vote choice, were obtained from the pre-election wave. Using split samples, ANES had two different versions of questionnaires named old and new, and each version included six policy issues designed to gauge respondents’ issue perceptions and attitudes. While three issues (i.e., government spending and services, defense spending, and assistance to blacks) were commonly included in both questionnaires, three different issues were asked in the two versions. In addition to the three common issues, the old version had three other issue items: medical insurance, job guarantee, and environment/job tradeoff. The new version included questions regarding universal health coverage, illegal immigrants’ citizenship process, and lower emission standards. For each issue, ANES asked four issue-related questions: respondents’ perception of issue importance, issue preferences, and perceptions of Barack Obama’s and John McCain’s issue positions.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 693

Measures Issue importance. Respondents’ perceptions of issue importance were measured by the question, “How important is this issue to you personally?” on a five-point scale ranging from “not important at all” (1) to “extremely important” (5). Responses to the six questions were added to create a composite index (α = .72). Inter-item reliability of issue importance was lower than the ideal level, which may be attributed to the presence of the issue publics who attend to a single or only a few issues instead of considering a broad range of issues. Issue publics would give high importance to certain issues of interest to them, whereas they would assign low importance to the remaining issues. This tendency is expected to function as a reduction of overall inter-item reliability for issue importance.

Issue knowledge. Respondents’ issue knowledge was measured by the relatively correct placement of the two presidential candidates’ issue positions using questions on a seven-point scale. For example, when a respondent placed Obama’s position on service spending closer to “government should provide more services” than McCain’s on the scale, the response was counted as a correct answer (1 = correct, 0 = incorrect). A six-item index, ranging from 0 to 6 points, was constructed by combining answers to the six issues (α = .66). Inter-item reliability of issue knowledge was marginal. In general, measures of political knowledge tend to have relatively low inter-item reliability because of measurement errors arising from guessing effects and variation in difficulty level among different items. First, some uninformed respondents may correctly answer knowledge questions by chance. The measurement error due to the guessing becomes more acute when knowledge questions employ multiple-choice items rather than open-ended items. Second, a battery of knowledge questions generally includes items having different levels of difficulty. Hard questions are useful to discriminate respondents’ knowledge levels, but they lessen overall reliability.38 As a specific domain of political knowledge, our issue-knowledge items should share the same general problems inherent in knowledge-related measurement, which might have caused marginal inter-item reliability.

Issue proximity. To measure respondents’ perceptions of the closeness between their own issue preferences and the two candidates’ positions, we used a scale developed in previous studies.39 In addition to questions about issue placement of the candidates, ANES also asked respondents to situate themselves on the same seven-point scale. The triangular relationships of the three questions determining respondents’ own issue preferences and perceptions of the two candidates’ issue positions generated the dis- tances for computation. First, for each issue, the distance between a respondent’s own preference and the perception of McCain’s issue positions was calculated. Distance between the respondent’s and Obama’s position was also computed using the same method. Second, the absolute distance from McCain was subtracted from the absolute distance from Obama. Each value implied strength as well as direction. For example, if a respondent

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 694 Journalism & Mass Communication Quarterly 91(4) put himself or herself at 2, McCain at 3, and Obama at 7 on the defense-spending issue scale, the proximity score for the issue was 4. Because a longer distance indicates a higher level of disagreement, a positive sign after subtraction means that the respondent concurred more with McCain than Obama. Conversely, a negative sign indicates that the respondent is closer to Obama than McCain in terms of issue proximity. Zero means the respondent does not see any difference between the two candidates’ issue stances. The possible range of each issue score was −6 to 6. Final issue-proximity scores were an index adding the proximity scores of the six issues (α = .78).

News attention. Attention to TV, newspaper, and the Internet was measured by questions determining respondents’ attention to campaign news or information from each source. Respondents answered on a five-point scale from “none” (1) to “a great deal” (5).

Vote choice. In postelection interviews, respondents indicated whom they voted for. Vote choice was dummy coded (McCain = 1, Obama = 0).

Control variables. Four demographic variables were added for control: gender (female = 57%), age (M = 47.32 years old, SD = 17.16), education (M = 13.13 years, SD = 2.57), and income (median = $30,000-$34,999). These variables have been found to influence political participation40 and news use.41 Education, especially, is a strong predictor of news attention as well as issue knowledge.42 Two political variables were also included: partisanship and campaign interest. Partisanship has been found to influence political attitudes and behaviors, including voting.43 Party identification was separated into Democrat (42.1%), Republican (18.9%), and Independent (37.5%). Campaign interest was controlled because greater interest in the presidential campaign would encourage people to pay more attention to news, which could affect the subsequent issue-voting process. Campaign interest was a single-item question asking how much a respondent was interested in the campaign.

Analytic Procedures Structural Equation Modeling (SEM) was used to test the hypothesized relationships among news attention, issue importance, issue knowledge, and issue proximity. For the analysis, respondents’ original issue-proximity scores (−6 to +6) were transformed into absolute values (0 to 6) because the related hypotheses concern the strength of issue proximity, not the direction of favorability toward either Obama or McCain. Vote choice was not included because the variable involved the direction of respondents’ selection of one candidate over the other. We cannot expect, for example, that greater news attention or issue knowledge would predict voting for either Obama or McCain. For the research question on vote choice, hierarchical logistic regression was conducted. We entered the original measurement of issue proximity, including the direction of favorability toward either one of the candidates, as a main predictor of vote choice. Demographic variables, partisanship, and campaign interest were controlled in all analyses.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 695

Table 1. Descriptive Statistics of Issue-Related Variables.

Importance Knowledge Proximity

Percentage of M of absolute M correct answers value Spending and services 3.66 61.1 1.70 Defense spending 3.54 63.6 1.58 Government medical 4.03 61.3 1.88 insurance Universal health 3.73 62.8 2.20 coverage Job guarantee 3.72 61.7 1.65 Illegal immigrants’ 3.20 40.7 1.15 citizenship Assistance to blacks 3.22 38.9 1.34 Environment vs. jobs 3.41 39.1 1.00 Lower emission 3.29 15.8 0.98 standards

Results Preliminary aggregate-level analysis showed that issue importance, issue knowledge, and issue proximity were associated with each other. Rank-order correlation indicated significant relationships between importance and knowledge (ρ = .67, p < .05), importance and proximity (ρ = .83, p < .01), and knowledge and proximity (ρ = .73, p < .05).44 The levels of respondents’ issue knowledge, perceptions of issue importance, and proximity varied among the six issue items (see Table 1). Figure 1 presents the results of SEM analysis controlling for six exogenous variables. There was one latent variable, news attention to traditional media, comprised of two indicators—newspaper and TV news attention. The model χ2 (60.35, df = 22) was significant at the .001 level, yet other statistics presented good fits: the goodness of fit index (GFI) of 1.00, the comparative fit index (CFI) of .99, the root mean square error of approximation (RMSEA) of .03, and the standardized root mean square residual (SRMR) of .02. The indices met one of the criteria of RMSEA < .06 and SRMR < .10.45 The normed chi-square, value of χ2 divided by degrees of freedom, of the model was 2.74, and it also indicated the model was acceptable.46 Squared multiple correlation of issue proximity was .46, indicating 46% of variance of the variable was explained in the model. Among eight direct paths that included main variables of news attention, issue importance, issue knowledge, and issue proximity, six were statistically significant at the .001 level. H1 predicted that news attention would be positively related to issue importance. While attention to traditional media was a significant predictor of issue importance, attention to campaign information on the Internet was not. The standardized coefficient of the path from traditional media and the Internet to issue importance was .23

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 696 Journalism & Mass Communication Quarterly 91(4)

.15***

.65*** TV .23*** TradionalT Issue Importance Media .16*** Newspaper .48*** .29*** Issue Proximitya .03 .53*** Internet Issue Knowledge .08***

.03

Figure 1. Result of hypothesized model. Note. Model Fit: χ2 (22) = 60.35, p < .001, GFI = 1.00, CFI = .99, RMSEA = .03, SRMR = .02. Solid lines indicate statistical significance of direct effects. Exogenous variables (gender, age, education, income, partisanship, and campaign interest) were entered as covariates. GFI = goodness of fit index; CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. aAbsolute value of issue proximity. Standardized coefficients, ***p < .001.

(p < .001) and .03 (p = .15), respectively. H1 was partially supported. H2 posited that news attention would be positively related to issue knowledge. As expected, all media variables were significant: direct effects of attention to traditional media (.29) and the Internet (.08) on issue knowledge were both statistically significant at the .001 level. Paying greater attention to news media contributed to greater knowledge about policy issues. H2 was supported. The next hypotheses concerned predictors of issue proximity. First, direct effects of both issue importance (.16, p < .001) and issue knowledge (.53, p < .001) on issue proximity were positive and significant. Therefore, H3 and H4 were confirmed. H5, predicting the relationship between news attention and issue proximity, yielded mixed results. Attention to traditional news (.15, p < .001) directly affected increase in issue proximity. The path from attention to the Internet news to issue proximity was not significant (.03, p = .15). Effects of decomposition, however, demonstrated that both traditional media and the Internet news attention exerted significant indirect effects on issue proximity through issue importance and issue knowledge. Indirect effects of attention on traditional media (.19) and the Internet (.05) on issue proximity were significant at the .001 level (see Table 2). Finally, we asked one research question about whether issue proximity had a stronger association with voting decisions among political independents who pay closer attention to news than others. To compare the effect sizes of issue proximity on vote choice depending on party identification and levels of news media attention,

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 697

Table 2. Effects Decomposition of Main Variables.

Endogenous variables

Issue Issue Issue Causal variables Effects importance knowledge proximity Traditional media Direct .23*** .29*** .15*** Indirect .19*** Total .23*** .29*** .34*** Internet Direct .03 .08*** .03 Indirect .05*** Total .03 .08*** .07*** Issue importance Direct .16*** Indirect Total .16*** Issue knowledge Direct .53*** Indirect Total .53***

Note. Bootstrapping was used to test significance of the standardized coefficients. ***p < .001.

respondents were first classified into three groups of Democrats, Republicans, and Independents. Those groups were divided again into high/low attention groups, using the median value of media attention as a cutoff point. As a result, a total of eighteen groups (three party identifications × two levels of attention × three media types) were created. Each hierarchical logistic regression analysis had two blocks: control block (age, gender, education, income, campaign interest, issue importance, and issue knowledge) and issue proximity. Because the research question focused on issue proximity, Table 3 presents the statistical values related only to issue proximity. As Table 3 shows, issue proximity was a significant predictor of vote choice at least at the .01 level in all eighteen analyses, regardless of party identification or news- attention levels. The first noteworthy point in the results is that the Independent group showed stronger association between issue proximity and vote choice than the Democrats and Republicans across different media types and attention levels. In the case of low-TV news-attention groups, for example, the Wald statistic of the Independent group (W = 22.67, p < .001) was much higher than for the Democratic (W = 8.27, p < .01) or Republican groups (W = 6.58, p < .01). In sum, voters who did not express identified partisanship relied more on issue proximity in choosing their next president. Comparisons between high and low news-attention groups yielded another notable finding. There were higher positive associations between issue proximity and vote choice among high news-attention groups than among low-attention groups in general. That is, issue proximity worked as a relatively stronger predictor in the high news- attention groups than in the low-attention groups, regardless of respondents’

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 698 Journalism & Mass Communication Quarterly 91(4)

Table 3. Issue Proximity as a Predictor of Vote Choice Depending on News-Attention Levels and Partisanship.

Low attention to news High attention to news

R2 R2 Exp (Nagelkerke’s) Exp (Nagelkerke’s) Walda (B) changeb Wald (B) change Democrat TV 8.27** 1.19 .14 24.07*** 1.2 .18 Newspaper 14.42*** 1.20 .15 19.16*** 1.19 .18 Internet 14.16*** 1.18 .13 15.84*** 1.18 .17 Independent TV 22.67*** 1.29 .38 61.98*** 1.33 .51 Newspaper 36.23*** 1.29 .43 45.78*** 1.35 .51 Internet 31.12*** 1.30 .35 48.29*** 1.35 .59 Republican TV 6.58** 1.56 .39 19.61*** 1.35 .30 Newspaper 9.86** 2.19 .48 10.65** 1.21 .19 Internet 11.35** 1.30 .26 14.86*** 1.51 .38

Note. Bold letters indicate higher values in comparisons of issue-proximity scores between low- and high- attention groups. Hierarchical logistic regression analyses, vote choice: 0 = Obama, 1 = McCain. aTest of statistical significance of the issue-proximity coefficient in the model. bAmount of variance explained by issue proximity in addition to amount by age, gender, education, income, campaign interest, issue importance, and issue knowledge. *p < .05. **p < .01. ***p < .001. partisanship. More importantly, the analyses also demonstrated differences in interaction between partisanship and levels of news attention: compared with partisan groups, the Independents had a consistently stronger relationship between issue proximity and vote choice among the high-attention group than the low-attention group, regardless of media type. To highlight this comparison, Table 3 marked a higher value of statistics between low- and high-attention groups by different partisanship groups, as shown in the bold numbers. These findings indicated that news attention had a stronger relationship with vote choice among nonpartisan voters through issue proximity. When respondents lacked partisan loyalties, either Democratic or Republican, they relied more on issues to determine for whom to vote. More importantly, the association between issue proximity and vote choice was stronger among those independents who paid greater attention to campaign news and information than did other groups.

Discussion This study illustrates that an individual’s news use plays an important role in the issue- voting process. Following the traditional view of issue voting, we propose that issue voters should have issue concerns, issue knowledge, and comparisons of their own

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 699 issue preferences against competing candidates’ policy stances, which consequently should lead them to vote for a candidate whose issue positions agree more closely with their own. SEM analysis demonstrates that attention to traditional news media, consisting of newspaper and TV, relates to these necessary conditions for issue voting. Those who pay greater attention to traditional news media tend to assign higher importance to policy issues and to have higher levels of issue knowledge. News attention also helps individuals perceive they are closer to one candidate in terms of issue positions, while they feel more distant from the other candidate. Simply, news attention contributes to voters’ differentiation of alternative policy positions, as espoused by the two different presidential candidates. Attention to campaign information on the Internet, however, had a much weaker association with the issue-voting process. While contributing to issue knowledge, attention to online information did not increase issue importance and proximity directly. Instead, attention to online information worked on issue proximity through issue importance and knowledge. This relatively weak association may result from a single-item measure, which would encompass attention to many different types of information sources, such as traditional news media websites, candidate websites, and even citizen-generated content. Another interpretation is that the Internet might be less effective than traditional news media in facilitating the electorate to engage in issue voting based on the consideration of a broad range of issues. Because the Internet allows greater individual selection of information, some people focus on issues of interest and bypass others they may come across when consuming traditional news media. The nonlinear style of information presentation on the websites makes it easier for the audience to avoid certain topics and issues, compared with the linear structure of traditional news media guiding the audience through spatial and time ordering of content.47 User selectivity and the organizing structure of information on the web may then enhance concerns and knowledge about issues of personal interest, but they may decrease the opportunity for people to encounter, learn, and form opinions about diverse issues. Another important finding of the current study is the different relationships between news and political information use and issue proximity among political independents and partisans. When respondents were divided into separate groups according to their news-attention levels and party identification, the relationship between proximity and voting choice was stronger among political independents highly attentive to news and campaign information. The results seem to suggest that news use, as a source of issue- relevant information, has greater association with issue voting for independents. Overall, this study makes a unique contribution to research on the role of news media in issue voting. Previous studies typically failed to consider fully whether news use guides issue voting as a multiplicative process and how news use relates to voters’ evaluations of agreements between their own policy preferences and candidates’ stances. As an important exception, Kim and colleagues’ study demonstrated that news use relates to some key elements of issue voting: issue opinionation, issue knowledge, and issue consideration for candidate evaluation.48 Their operationalization of issue

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 700 Journalism & Mass Communication Quarterly 91(4) voting, as being the number of issues mentioned for candidate evaluation, however, does not adequately address whether voters’ assessments of issue congruence affect their vote choices, a question regarded as an essential element of issue voting.49 As one of the first studies exploring the association between news use and assessment of issue agreement, the current research documents the idea that news use not only enhances general issue consideration for candidate evaluation, but also facilitates evaluation of issue congruence, which ultimately converts to candidate choice. Also, this study contributes to the area of issue-voting research by examining how partisanship interacts with news use. Issue voting has been considered as a function of erosion of partisanship and growing availability of campaign information via news media.50 This observation suggests issue voting would be more common among political independents who greatly rely on news media for campaign information. Our study provides rare empirical evidence supportive of this argument. The current research has specific implications for journalists, candidates, and campaign professionals. The significant role of news media for issue voting found in this study encourages journalists and news organizations to devote coverage to substantive issues, rather than candidates’ personal characteristics and horserace coverage. In doing so, journalists will fulfill their vital roles of helping the public make rational, informed decisions in elections, as well as mitigating public cynicism and distrust toward the news media and politics.51 Our study also reminds candidates that emphasizing their policy stands can be the most effective strategy to earn votes from political independents. When targeting independents, candidates may consider tailoring their campaign messages to differentiate their issue positions from those of their competitors. Limitations of the current study concern an assumed causal direction between issue proximity and vote choice. Although we found that voters tend to choose a candidate whose issue positions are closer to their own, it does not necessarily mean that voters’ evaluations about issue positions determine their vote choices. Instead, people may rationalize their issue preferences after deciding for whom they will vote. The direction of influence from news use to issue-related variables presumed in this study raises similar causality concerns. Following prior research, we viewed news attention as an antecedent or predictor of issue importance, knowledge, and proximity. In fact, studies using panel data found that news use precedes issue-specific knowledge,52 as well as general political knowledge.53 Also, many analyses in the agenda- setting tradition have documented effects of news use on perception of issue importance based on both experimental54 and longitudinal data.55 Even though these prior investigations provide theoretical rationales for the causal inference made in this study, the causality issue has yet to be addressed fully. We admit that reverse causation is possible in the direction of effects running from issue importance, knowledge, and proximity to news use. People who assign greater importance to issues and have greater issue knowledge may be more likely to consume news media to seek further information about the issues and reinforce their issue attitudes. The cross-sectional data we employed, therefore, do not allow us to draw a firm conclusion that news attention leads to the issue-voting process. Another drawback relates to the measurement of concepts included in this study.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 701

Due to the inherent limitations of secondary data, some of the measurements used in this study were less than ideal to tap the concepts of our interest. First, single-item measures of news attention raise concerns about reliability, as well as validity. Compared with multi-item scales, single-item scales tend to be more prone to measurement error, and at the same time may not fully capture the complexity of a construct. Communication scholars have debated what constitutes the valid measurement of news use to grasp its full influence on political cognitions, attitudes, and behaviors. Some researchers note the relatively weak predictive power of self-reported measures of news use for actual news reception, which makes regarding news attention as a meaningful predictor of the political effects of news a questionable practice.56 Other investigators identified distinct effects of different dimensions of news use, such as exposure to specific content, reliance, and gratification seeking.57 Our news-attention measurements, based on single items, neither tap fully into attention to specific news content nor consider other dimensions of news use that could relate to the issue-voting process. Especially, considering the fragmented and diversified news environment of recent election campaigns, questions asking about respondents’ general news use may not properly tap into their habits. Attention to different news outlets and programs may highlight distinctive influences on the issue-voting process. Second, the Internet attention measure obtained from the post-wave survey raises concerns about the causal inference assumed in this study. The retrospective nature of the question regarding Internet attention reduces the validity worry because the question inquired about respondents’ past attention during the campaign rather than present attention after the election. It should be noted, however, that retrospective questions tend to undermine reliability of measurement. Respondents’ memory decay and extra effort to recall past behaviors might have led to underreporting or overreporting of their actual Internet attention. We expect future research would continue to elaborate on the role of news media in the issue-voting process. In our study, we looked at people’s overall levels of news attention, knowledge, and perceptions about political issues. Not all issues will be equally considered when individuals make their voting decisions, though. Issue publics, for example, should give more consideration to certain issues over others. They may employ news media and sources differently to satisfy their parochial information goals, which could be hard to achieve from traditional news media that cater to a general audience. Also, issue voting may not necessarily represent vote choice based on the electorates’ rational calculations of candidates’ competing issue stances. Some issue voters may choose a candidate whose issue positions apparently contradict their own personal or group interests.58 We propose that subsequent studies should consider how diverse groups deploy different news outlets and content to engage in potentially different types of issue voting.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 702 Journalism & Mass Communication Quarterly 91(4)

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. See, for example, Bernard R. Berelson, Paul F. Lazarsfeld, and William N. McPhee, Voting: A Study of Opinion Formation in a Presidential Election (Chicago: University of Chicago Press, 1954); Norman H. Nie, Sidney Verba, and John R. Petrocik, The Changing American Voter (Cambridge, MA: Harvard University Press, 1976). 2. Michael X. Delli Carpini and Scott Keeter, What Americans Know about Politics and Why It Matters (New Haven, CT: Yale University Press, 1996); Paul Goren, “Political Expertise and Issue Voting in Presidential Elections,” Political Research Quarterly 50 (June 1997): 387-412. 3. Ole Borre, Issue Voting: An Introduction (Langelandsgade, Denmark: Aarhus University Press, 2011); Russell J. Dalton, Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies (Washington, DC: CQ Press, 2008); Arthur H. Miller, Warren E. Miller, Alden S. Raine, and Thad A. Brown, “A Majority Party in Disarray: Policy Polarization in the 1972 Election,” American Political Science Review 70 (September 1976): 753-78. 4. Pew Research Center, American Values Survey: Section 9, Trends in Party Affiliation (Washington, DC: Pew Research Center, 2012), http://www.people-press.org/2012/06/04/ section-9-trends-in-party-affiliation/ (accessed February 2, 2014). 5. Alan I. Abramowitz and Kyle L. Saunders, “Is Polarization a Myth?” Journal of Politics 70 (April 2008): 542-55; Gary C. Jacobson, “Polarized Politics and 2004 Congressional and Presidential Elections,” Political Science Quarterly 120 (summer 2005): 199-218. 6. Shanto Iyengar and Kyu S. Hahn, “Red Media, Blue Media: Evidence of Ideological Selectivity in Media Use,” Journal of Communication 59 (March 2009): 19-39; Michael Pfau, Brian Houston, and Shane M. Semmler, Mediating the Vote: The Changing Media Landscape in U.S. Presidential Campaigns (Lanham, MD: Rowman & Littlefield, 2007). 7. Ester de Waal and Klaus Schoenbach, “Presentation Style and Beyond: How Print Newspapers and Online News Expand Awareness of Public Affairs Issues,” Mass Communication and Society 11 (2, 2008): 161-76; Maxwell E. McCombs and Donald Lewis Shaw, “The Agenda-Setting Function of Mass Media,” Public Opinion Quarterly 36 (summer 1972): 176-87. 8. Steven H. Chaffee, Xinshu Zhao, and Glenn Leshner, “Political Knowledge and the Campaign Media of 1992,” Communication Research 21 (June 1994): 305-24; Sei-Hill Kim, Dietram A. Scheufele, and James Shanahan, “Who Cares about the Issues? Issue Voting and the Role of News Media during the 2000 U.S. Presidential Election,” Journal of Communication 55 (March 2005): 103-21. 9. Shanto Iyengar and Donald R. Kinder, News That Matters: Television and American Opinion (Chicago: University of Chicago Press, 1987); Marilyn S. Roberts, “Predicting Voting Behavior via the Agenda-Setting Tradition,” Journalism Quarterly 69 (winter 1992): 878-92; Marilyn S. Roberts, “Political Advertising’s Influence on News, the Public and Their Behavior,” in Communication and Democracy: Exploring the Intellectual Frontiers in Agenda-Setting Theory, ed. Maxwell E. McCombs, Donald Lewis Shaw, and David Hugh Weaver (Mahwah, NJ: Lawrence Erlbaum, 1997), 85-96; Tamir Sheafer and Gabriel Weimann, “Agenda Building, Agenda Setting, Priming, Individual Voting Intentions, and

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 703

the Aggregate Results: An Analysis of Four Israeli Elections,” Journal of Communication 55 (June 2005): 347-65; Denis Wu and Renita Coleman, “Advancing Agenda-Setting Theory: The Comparative Strength and New Contingent Conditions of the Two Levels of Agenda-Setting Effects,” Journalism & Mass Communication Quarterly 86 (winter 2009): 775-89. 10. Richard A. Brody and Benjamin I. Page, “Comment: The Assessment of Policy Voting,” American Political Science Review 66 (June 1972): 450-58; Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes, The American Voter (NY: John Wiley, 1960); Edward G. Carmines and James A. Stimson, “The Two Faces of Issue Voting,” American Political Science Review 74 (March 1980): 78-91; Anthony Downs, An Economic Theory of Democracy (NY: Harper & Row, 1957). 11. Berelson, Lazarsfeld, and McPhee, Voting: A Study of Opinion Formation; Campbell et al., The American Voter; Philip E. Converse, “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. David Apter (NY: Free Press, 1964), 206-61. 12. Bryan Caplan, The Myth of the Rational Voter: Why Democracies Choose Bad Policies (Princeton, NJ: Princeton University Press, 2007); Delli Carpini and Keeter, What Americans Know about Politics. 13. Paul R. Abramson, John Herbert Aldrich, and David W. Rhode, Change and Continuity in the 2004 Elections (Washington, DC: CQ Press, 2005); Carmines and Stimson, “The Two Faces of Issue Voting.” 14. McCombs and Shaw, “The Agenda-Setting Function of Mass Media.” 15. Iyengar and Kinder, News That Matters. 16. Roberts, “Predicting Voting Behavior via the Agenda-Setting Tradition”; Roberts, “Political Advertising’s Influence on News”; Wu and Coleman, “Advancing Agenda- Setting Theory.” 17. Chaffee, Zhao, and Leshner, “Political Knowledge and the Campaign Media of 1992”; Glenn J. Hansen and William L. Benoit, “Communication Forms as Predictors of Issue Knowledge in Presidential Campaigns: A Meta-analytic Assessment,” Mass Communication and Society 10 (2, 2007): 189-210. 18. De Waal and Schoenbach, “Presentation Style and Beyond.” 19. Kim, Scheufele, and Shanahan, “Who Cares about the Issues?” 20. For a review, see Maxwell E. McCombs, Setting the Agenda: The Mass Media and Public Opinion (Malden, MA: Polity Press, 2004). 21. Jon A. Krosnick, “Government Policy and Citizen Passion: A Study of Issue Publics in Contemporary America,” Political Behavior 12 (March 1990): 59-92; Kathleen M. McGraw, Milton Lodge, and Patrick Stroh, “On-line Processing in Candidate Evaluation: The Effects of Issue Order, Issue Salience, and Sophistication,” Political Behavior 12 (March 1990): 41-58. 22. Penny S. Visser, George Y. Bizer, and Jon A. Krosnick, “Exploring the Latent Structure of Strength-Related Attitude Attributes,” in Advances in Experimental Social Psychology, ed. Mark P. Zanna (San Diego, CA: Academic Press, 2006), 5. 23. John H. Aldrich, John L. Sullivan, and Eugene Borgida, “Foreign Affairs and Issue Voting: Do Presidential Candidates ‘Waltz before a Blind Audience?’” American Political Science Review 83 (March 1989): 123-41; Krosnick, “Government Policy and Citizen Passion.” 24. McCombs and Shaw, “The Agenda-Setting Function of Mass Media.” 25. De Waal and Schoenbach, “Presentation Style and Beyond.” 26. Thomas E. Patterson and Robert D. McClure, The Unseeing Eye: The Myth of Television Power in National Politics (NY: G. P. Putnam’s, 1976).

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 704 Journalism & Mass Communication Quarterly 91(4)

27. Hansen and Benoit, “Communication Forms as Predictors of Issue Knowledge.” 28. Borre, Issue Voting; Downs, An Economic Theory of Democracy; George Rabinowitz and Stuart Elaine Macdonald, “A Directional Theory of Issue Voting,” American Political Science Review 83 (March 1989): 93-121. 29. Krosnick, “Government Policy and Citizen Passion”; Richard R. Lau and David P. Redlawsk, “Voting Correctly,” American Political Science Review 91 (September 1997): 585-98; George Rabinowitz, James W. Prothro, and William Jacoby, “Salience as a Factor in the Impact of Issues on Candidate Evaluation,” Journal of Politics 44 (February 1982): 41-63; Penny S. Visser, Jon A. Krosnick, and Joseph P. Simmons, “Distinguishing the Cognitive and Behavioral Consequences of Attitude Importance and Certainty: A New Approach to Testing the Common-Factor Hypothesis,” Journal of Experimental Social Psychology 39 (March 2003): 118-41. 30. Iyengar and Kinder, News That Matters. 31. Patrick Fournier, André Blais, Richard Nadeau, Elisabeth Gidengil, and Neil Nevitte, “Issue Importance and Performance Voting,” Political Behavior 25 (March 2003): 51-67; Jon A. Krosnick, “The Role of Attitude Importance in Social Evaluation: A Study of Policy Preferences, Presidential Candidate Evaluations, and Voting Behavior,” Journal of Personality and Social Psychology 55 (August 1988): 196-210; Rabinowitz, Prothro, and Jacoby, “Salience as a Factor”; David E. RePass, “Issue Salience and Party Choice,” American Political Science Review 65 (June 1971): 389-400. 32. Delli Carpini and Keeter, What Americans Know about Politics; Goren, “Political Expertise and Issue Voting.” 33. Stephen P. Nicholson, Adrian Pantoja, and Gary M. Segura, “Political Knowledge and Issue Voting among the Latino Electorate,” Political Research Quarterly 59 (June 2006): 259-71. 34. Rabinowitz and Macdonald, “A Directional Theory of Issue Voting.” 35. Jon A. Krosnick, David S. Boninger, Yao C. Chuang, Matthew K. Berent, and Catherine G. Carnot, “Attitude Strength: One Construct or Many Related Constructs?” Journal of Personality and Social Psychology 65 (December 1993): 1132-51; Abraham Tesser, Leonard Martin, and Marilyn Mendolia, “The Impact of Thought on Attitude Extremity and Attitude-Behavior Consistency,” in Attitude Strength: Antecedents and Consequences, ed. Richard E. Petty and Jon A. Krosnick (Mahwah, NJ: Lawrence Erlbaum, 1995), 1-24. 36. Dalton, Citizen Politics. 37. Abramowitz and Saunders, “Is Polarization a Myth?”; Jacobson, “Polarized Politics and 2004 Congressional and Presidential Elections.” 38. Delli Carpini and Keeter, What Americans Know about Politics; Jeffrey J. Mondak, “Developing Valid Knowledge Scales,” American Journal of Political Science 45 (January 2001): 224-38; Jeffrey J. Mondak, “Political Knowledge and Cross-National Research on Support for Democracy” (paper presented at the Latin American Public Opinion Project-United Nations Development Program [LAPOP-UNDP] Workshop on Candidate Indicators for the UNDP Democracy Support Index [DSI], Nashville, TN, May 5-6, 2006). 39. Rabinowitz, Prothro, and Jacoby, “Salience as a Factor”; Visser, Krosnick, and Simmons, “Distinguishing the Cognitive and Behavioral Consequences.” 40. Dalton, Citizen Politics. 41. McCombs, Setting the Agenda. 42. Chaffee, Zhao, and Leshner, “Political Knowledge and the Campaign Media of 1992.” 43. Dalton, Citizen Politics.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Hyun and Moon 705

44. Additional Pearson’s r analyses also found the significant correlations between issue importance and issue knowledge (r = .75, p < .05), issue importance and issue proximity (r = .80, p < .05), and finally, issue knowledge and issue proximity (r = .84, p < .01). 45. Li-tze Hu and Peter M. Bentler, “Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria versus New Alternatives,” Structural Equation Modeling: A Multidisciplinary Journal 6 (1, 1999): 1-55. 46. Kenneth A. Bollen, Structural Equations with Latent Variables (NY: Wiley, 1998). 47. De Waal and Schoenbach, “Presentation Style and Beyond”; Markus Prior, “News vs. Entertainment: How Increasing Media Choice Widens Gaps in Political Knowledge and Turnout,” American Journal of Political Science 49 (July 2005): 577-92; David Tewksbury and Scott L. Althaus, “Differences in Knowledge Acquisition among Readers of the Paper and Online Versions of a National Newspaper,” Journalism & Mass Communication Quarterly 77 (fall 2000): 457-79. 48. Kim, Scheufele, and Shanahan, “Who Cares about the Issues?” 49. Berelson, Lazarsfeld, and McPhee, Voting: A Study of Opinion Formation in a Presidential Election; Brody and Page, “Comment: The Assessment of Policy Voting”; Campbell et al., The American Voter; Downs, An Economic Theory of Democracy; Kenneth J. Meier and James E. Campbell, “Issue Voting: An Empirical Evaluation of Individually Necessary and Jointly Sufficient Conditions,” American Political Quarterly 7 (January 1979): 21-50; Nie, Verba, and Petrocik, The Changing American Voter. 50. Dalton, Citizen Politics. 51. Joseph N. Cappella and Kathleen Hall Jamieson, Spiral of Cynicism: The Press and the Public Good (NY: Oxford University Press, 1997). 52. See, for example, Steven H. Chaffee and Joan Schleuder, “Measurement and Effects of Attention to Media News,” Human Communication Research 13 (September 1986): 76-107. 53. William P. Eveland, Andrew F. Hayes, Dhavan V. Shah, and Nojin Kwak, “Understanding the Relationship between Communication and Political Knowledge: A Model Comparison Approach Using Panel Data,” Political Communication 22 (4, 2005): 423-46. 54. See, for example, Iyengar and Kinder, News That Matters. 55. See, for example, Hans-Bernd Brosius and Hans Mathias Kepplinger, “The Agenda-Setting Function of Television News: Static and Dynamic Views,” Communication Research 17 (April 1990): 183-211. 56. Vincent Price and John Zaller, “Who Gets the News? Alternative Measures of News Reception and Their Implications for Research,” Public Opinion Quarterly 57 (summer 1993): 133-64. 57. Jack M. McLeod and Daniel G. McDonald, “Beyond Simple Exposure: Media Orientations and Their Impact on Political Processes,” Communication Research 12 (January 1985): 3-34. 58. See, for example, Thomas Frank, What’s the Matter with Kansas? (NY: Henry Holton, 2004).

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 JMQXXX10.1177/1077699014550088Journalism & Mass Communication QuarterlyLee and Chyi 550088research-article2014

Audience Motivations and Responses

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 706–724 Motivational Consumption © 2014 AEJMC Reprints and permissions: Model: Exploring the sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550088 Psychological Structure of jmcq.sagepub.com News Use

Angela M. Lee1 and Hsiang Iris Chyi2

Abstract An informed citizenry is important to the functioning of democratic political systems. Despite the proliferation of news outlets over the past decade, news consumption among the general American public remains shallow, and an empirical understanding of why people consume news remains limited. Integrating existing theories, this study proposes the Motivational Consumption Model, which explains why age and education predict news consumption through three psychological mediators: news beliefs, news motivations, and news attitudes. Using a national survey (N = 1,143) and structural equation modeling, this study reveals the psycho-behavioral workings of news consumption.

Keywords news consumption, behavior prediction, Motivational Consumption Model, structural equation modeling

An informed, vigilant, and active citizenship is essential to the health and functioning of a democratic political system, and to a large extent, this revolves around the issue of news consumption.1 Any democratic political system is grounded on the premise that its citizens participate actively in the political process, which includes becoming informed on important issues and making rational, informed political decisions.2 News comprises raw material from which public opinions are formed, and it also alerts the public to information needed to make political decisions (i.e., whom to vote for or

1University of Texas at Dallas, Dallas, TX, USA 2University of Texas at Austin, Austin, TX, USA Corresponding Author: Angela M. Lee, University of Texas at Dallas, 800 West Campbell Road, Mail Station AT10, Richardson, TX75080, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Downloaded from http://www.elearnica.ir Lee and Chyi 707

Figure 1. Motivational Consumption Model.

which propositions to endorse).3 A plethora of empirical studies on agenda setting, priming, and framing have demonstrated that the press has the ability to influence what issues we think about and how we think about them.4 Despite the rise of the Internet, cable television, and mobile devices, news consumption remains relatively shallow among the general American public. For instance, on average, news audiences across all four major news media (online, print, radio, and TV) spent only about eighteen minutes consuming the news yesterday,5 and 17% of Americans did not consume any news yesterday.6 The implication is important: Unless we find ways to encourage news consumption, which in turn promotes civic engagement and political participation, the future of American democracy may be jeopar- dized by a growing number of citizens who are neither politically informed nor engaged. The purpose of this study is to offer a psychological understanding of news consumption from a theory-driven approach—specifically, it seeks to explain why people consume news from a behavioral prediction perspective and, based on the findings, recommends how to encourage news consumption. In order to change people’s news consumption behavior, it is essential to understand what drives people to consume news. Drawing on the Reasoned Action Model (RAM) and the uses and gratifications paradigm, this study proposes the Motivational Consumption Model (see Figure 1), which targets psychological drivers of news consumption beyond what demographic factors can explain. Specifically, the Motivational Consumption Model uses news beliefs, news motivations, and news attitudes, key mediators between demographics and news consumption, to explain why those who are older and with higher education are more likely to consume news.7

Literature Review Demographics and News Consumption Demographic factors have long been of interest to the study of news consumption, including newspaper use8 and news use on the Internet,9 mobile devices,10 and social media.11 Among various factors, education and age positively influence consumption

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 708 Journalism & Mass Communication Quarterly 91(4) of news on politics, public affairs, and current events.12 Specifically, they are positively associated with consumption of news across television, radio, online, and print newspapers, as well as time spent consuming news in both digital and traditional platforms.13 While such findings are helpful in suggesting who is more likely to consume news at the aggregate level, analysis at this level does not explain why age and education positively influence news consumption. In fact, the lack of empirical examination on psycho-behavioral factors of news consumption remains an impediment that prevents news consumption prediction and thwarts development of interventions that promote news consumption.

Uses and Gratifications News consumption has most often been studied under the uses and gratifications framework, an audience-centric approach that seeks to understand how and why people use different media to satisfy different needs or gratifications. Studies find that audiences often use news to fulfill their need to know, to survey what is going on in the world, to keep up with the way the government performs its functions,14 to fulfill their need for entertainment and relaxation,15 for the purpose of opinion formation affirmation or avoidance,16 and for its social values (i.e., to have something to talk about with others).17 The literature is rich with different typologies, although a recent study has synthesized findings from an array of studies on news consumption motivations and reduced the different typologies down to four overarching motivations (information-driven, entertainment-driven, social-driven, and opinion-driven).18 While people have different motivations for consuming news, most use news primarily for informational gains.19

Behavioral Prediction and the Reasoned Action Model (RAM) The RAM is an updated combination of the Theory of Reasoned Action, Theory of Planned Behavior, the Health Belief Model, and Social Cognitive Theory.20 It offers a theoretical explanation for the ways in which different background factors influence behavior, allows for better understanding of various components that contribute to a given behavior, and is a robust theoretical framework for prediction across different behavioral domains.21 The RAM operates under the assumption that human social behavior is fundamentally influenced by people’s beliefs about the behavior under consideration, and that contrary to popular conception, human social behavior is determined by a finite number of key factors proposed in the model.22 In health communication, the primary domain in which the RAM flourishes, the model has been used to examine a variety of health-related behaviors, such as condom use, breast self-examination, smoking, and health screening.23 These studies suggest that the weight of influence of the three theoretical components—attitudes, norms, and self-efficacy—is behavior-specific, and that the success of behavioral interventions (e.g., those that promote a desirable behavior or discourage an undesirable behavior) is often determined by focusing on one of the three components that has the most impact on the behavior in question.24

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 709

The RAM contends that it is a comprehensive model that accounts for all necessary variables to predict behavior, and relegates all other variables related to a particular behavior to distal variables that are not of primary focus in the theory. This does not mean that all other variables are ignored. Instead, the RAM argues that all other variables are already accounted for, in some shape or form, in attitudes, social norms, and self-efficacy. For example, news access is measured under self-efficacy (i.e., whether one is able to consume news when one wants to), news interest is accounted for by attitudes (i.e., the more interesting news is the more favorable attitude one will have toward a news source), and social recommendations are part of social norms (i.e., the more one’s family and friends recommend a news site, the more likely one will visit a news site).

Predicting News Consumption: An Extended RAM Approach The RAM posits that attitudes, social norms, and self-efficacy will have different weights of influence on a behavior depending on the behavior in question.25 For example, while attitudes are found to carry the most weight for behaviors such as soft drink consumption and drug use among college students, self-efficacy is found to be the most influential in behaviors such as applying for a promotion, and social norms are the strongest predictors of behaviors such as driving after drinking.26 Notwithstanding the probable influence of perceived norms and self-efficacy on news consumption, this study centers on the effect of attitudes in influencing news consumption under the RAM framework. Particularly given the focus of this model on news consumption, attitudes are prioritized because perceived norms and self-efficacy are arguably less disputable, and less likely to be the defining factor in influencing news consumption. For instance, research indicates that a majority of Americans believe it is their responsibility to stay informed and engaged with news,27 which suggests a commonly held normative belief about the social desirability of news consumption. Other more controversial behaviors may involve different, if not conflicting, normative beliefs and considerations in relation to the behavior. For example, in terms of adolescent sexual behavior, research shows that whereas teenagers are more likely to believe that their parents disapprove of teenage sex, they are also more likely to believe that their friends and partners approve of premarital sex, and such normative contentions may have different behavioral influences.28 Moreover, given the relative ease of access to free online news content and the growing penetration of digital devices in recent years,29 it is reasonable to assume that news consumption is within most people’s capability, thus minimizing the role self- efficacy (e.g., the extent to which individuals can perform the behavior of interest) plays in contributing to, or preventing, news consumption. Furthermore, given the nature of news consumption, instead of focusing on the intention-behavior link as proposed by the RAM (e.g., the link between a verbal indication of wanting to carry out the behavior and actually carrying out the behavior), this study focuses on the mediating effects of news motivations (e.g., reasons for carrying out a certain behavior), which draws on the uses and gratifications literature.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 710 Journalism & Mass Communication Quarterly 91(4)

The RAM argues that every behavior is influenced by a different set of factors. For example, while running and swimming are both aerobic exercises, they are subject to different behavioral influences (i.e., running may be affected more by one’s attitudes toward running, whereas swimming may be more affected by self-efficacy factors such as distance between where one lives and the nearest public swimming pool). Similarly, running and not running should be considered two distinct behaviors and be studied differently. Following this logic, this model can only be generalized to news consumers, not non-consumers.

News beliefs. According to the RAM, behavioral beliefs are people’s perception of the attributes of a behavior in question, and are the basis on which attitudes toward a behavior are formed. The RAM contends that beliefs originate from a variety of sources that include personal experiences, formal education, media exposure, and interpersonal communication. Because individual differences such as demographic factors influence people’s exposure, retention, and interpretation of different experiences and information, the RAM presupposes that different demographic variables are likely to contribute to dissimilar beliefs, and thus assumes beliefs to be the causal consequence of demographic differences.30 In the context of news consumption, news beliefs are about people’s general view of the value of news, which encompasses beliefs about the importance of news in serving as a mobilizing agent that empowers the public. The Motivational Consumption Model posits that positive beliefs about the news influence news attitudes and news motivations in that the more people believe news to be important and empowering, the more likely people will possess favorable attitudes toward news, and the more likely people will be motivated to consume news.

News attitudes. Attitudes are innately and affectively evaluative on a unidimensional scale in that they generally lie on the continuum of two extremes, such as good and bad, and are determined by beliefs about the behavior that are the most salient.31 In other words, while most of us may have an infinite number of beliefs about a behavior that represent an array of behavioral attributes or characteristics, so long as the most important beliefs are captured, our general attitudes may be extrapolated. According to the Motivational Consumption Model, news attitudes are manifest in the perceived enjoyableness and advantages of news consumption, and it is theorized that the more favorable attitudes one has toward news, the more likely one will consume news. Nonetheless, the Motivational Consumption Model contends that news attitudes alone are necessary but insufficient in leading to news consumption, particularly if we take into consideration the “goal-orientedness” of media use from the uses and gratifications paradigm. Notwithstanding the importance of positive attitudes in leading to news consumption, the model suggests that audience-driven motivation to consume news is another determining factor that influences news consumption. Thus, in addition to the causal link between news beliefs and news attitudes, and following the RAM framework, the Motivational Consumption Model includes news motivations as an endogenous variable that is the consequence of news beliefs, and partial predictor of news attitudes.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 711

News motivations. Motivations have often been studied under the uses and gratifications paradigm to examine psychological drivers of media use,32 and are also used interchangeably with gratifications in the uses and gratifications literature.33 Uses and gratifications is an audience-centric approach that focuses on understanding why and how people use media, and it assumes media use is goal-oriented.34 The uses and gratifications literature has uncovered a number of motivations for media use. For example, research shows that while some people consume news for its social values or interpersonal utilities35; others consume news for relaxation, laughter, humor, or excitement36; and still others consume news to fulfill their need to know, for surveillance, or to keep up with the way the government performs its functions.37 In the context of the Motivational Consumption Model, the scope of news motivations is restricted to information-driven reasons for news consumption because the nature of news on politics, public affairs, and current events is predominantly to inform rather than to entertain or amuse (albeit it is possible that such news also entertains and amuses select news audiences). In sum, the Motivational Consumption Model posits the following interrelationships among news beliefs, news attitudes, and news motivations: (1) those who have positive news beliefs are more likely to have more favorable attitudes toward news, (2) those who have positive news beliefs are more likely to be motivated to consume news, (3) those who are motivated to consume news are more likely to have favorable attitudes toward news, and (4) those who have more favorable attitudes toward news and are more motivated to consume news are more likely to consume news.

News consumption. Recent empirical work has begun examining the multifaceted nature of news consumption to include not only demographic factors and news use, but also format preference and paying intent.38 While the latter two theoretical constructs are particularly important to studies of the economics of news consumption, they are outside of the scope of the Motivational Consumption Model. Instead, the Motivational Consumption Model centers on news use. News use and news consumption are thus used synonymously in this study to describe the frequency with which people seek out news on politics, public affairs, and current events.

Hypotheses The hypotheses tested in this study are implicit in the proposed structural model (see Figure 1) but also presented below:

H1: Age positively predicts news beliefs, controlling for education. H2: Education positively predicts news beliefs, controlling for age. H3: The more positive beliefs one has about news, the more favorable attitudes one will have toward news, controlling for news motivations. H4: The more positive beliefs one has about news, the more one will be motivated to consume news.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 712 Journalism & Mass Communication Quarterly 91(4)

H5: The more one is motivated to consume news, the more favorable attitudes one will have toward news, controlling for news beliefs. H6: The more favorable attitudes one has toward news, the more one will consume news, controlling for news motivations. H7: The more one is motivated to consume news, the more one will consume news, controlling for news attitudes.

To further examine the proposed mediation model, the following hypotheses regarding the indirect effects of demographic factors and news beliefs on news consumption are tested:

H8: The effect of age on news consumption is mediated by news beliefs, news motivations, and news attitudes in the structural model. H9: The effect of education on news consumption is mediated by news beliefs, news motivations, and news attitudes in the structural model. H10: The effect of news beliefs on news consumption is mediated by news motivations and news attitudes in the structural model.

Method To test the ten hypotheses, a national online panel survey was conducted between February 23 and 29, 2012, by the Office of Survey Research, which operates under the Annette Strauss Institute.39 The sample consisted of 1,143 adults (18+ years old), and it is representative of the U.S. adult population as reported in the most recent data from the U.S. Census Bureau.40 The completion rate is 66%.41 The size of this sample and its representativeness in terms of key demographic variables (age, gender, income, and race) contribute to this study’s reliability, stability, and generalizability.

Measurements Demographic predictors. Age and education are the exogenous variables in the structural model. Age is coded as an interval variable where one unit increase equates with one year older (M = 46.42, SD = 15.76), and education ranges from “some high school or less” to “master’s, MD, JD, or PhD,” with “some college” being the mode (35.1%).

News beliefs. News beliefs are operationalized as an index that centers on people’s general views about the value of news. News beliefs are the sum of the following two items that are based on a 7-point Likert-type scale ranging from strongly disagree (1) to strongly agree (7): How much do you agree or disagree with the following statements about the news in general? (a) News is important to you. (b) Being informed is empowering (range = 2-14, M = 11.28, SD = 2.66, Cronbach’s α = .75).42

News motivations. News motivations are operationalized as an index about consuming news for information purposes. News motivations comprise the sum of the following

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 713 five news consumption motivation items: most of us have different reasons and motivations for reading/watching/listening to the news. Thinking back to the sources you get your news from, with 1 being “strongly disagree” and 7 being “strongly agree,” please indicate the extent to which the following statements describe why you get news: You get news . . . (a) to find out what’s going on in the world, (b) to keep up with the way your government functions, (c) to make yourself an informed citizen, (d) to gain important new information, (e) to fulfill your need to know (range = 5-35, M = 28.33, SD = 6.03, Cronbach’s α = .89).

News attitudes. News attitudes are measured as an index comprising people’s dispositions to respond with some degree of favorableness toward news. Specifically, news attitudes are the sum of the following two items based on a 7-point Likert-type scale from strongly disagree (1) to strongly agree (7): How much do you agree or disagree with the following statements about the news in general? (a) Getting the news is enjoy- able to you. (b) Getting the news is advantageous to you (range = 2-14, M = 10.52, SD = 2.92, Cronbach’s α = .84).

News consumption. Consumption of news is measured by this question: “People have different preferences as to where they get their news. Thinking back to your own habits, how often do you read/watch/listen to the following news sources either in traditional (e.g., print) or digital (e.g., online) format, with (1) never, (2) seldom, (3) sometimes, (4) often: local newspapers, national newspapers (e.g., New York Times, USA Today, The Daily, CSM), CNN, MSNBC, FOX, network national news (e.g., Good Morning America, NBC Nightly News, CBS Evening News), local TV newscasts, cable Sunday talk shows (e.g., FOX News Sunday, State of the Union), network Sun- day talk shows (e.g., This Week, Meet the Press, State of the Union), Google News, and Yahoo News?” An index was created by summing the responses to these eleven questions (range = 11-44, M = 27.10, SD = 7.30, Cronbach’s α = .84).

Statistical Analysis To test the ten hypotheses, this study performed simultaneous maximum likelihood regression analyses under the structural equation modeling (SEM) framework using MPlus 6.1. This statistical approach allows for a holistic assessment of the proposed theoretical model, and offers a more intricate portrayal of all theorized direct and indirect effects among the exogenous and endogenous variables in the structural model.

Results Sample Overview Representative of the U.S. adult population, the sample used in this study compares favorably to the most recent U.S. census data on age, gender, income, and race. Twenty-eight percent of the respondents reported being between the ages of eighteen

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 714 Journalism & Mass Communication Quarterly 91(4) and thirty-four, 36% between thirty-five and fifty-four, and 35%, fifty-five or older. Nearly half (49%) of the respondents are male. Forty-nine percent of the respondents reported making less than $50,000 a year, 32% between $50,000 and $100,000 a year, and 19% more than $100,000 a year. More than two-thirds of the respondents (71%) are self-identified as white, 15% Hispanics, 15% African Americans, and 7% Asians.

Goodness of Fit Indices The fit of the model in Figure 1 was evaluated using a maximum likelihood algorithm. The model is statistically over-identified, which is desirable for theory-driven confir- matory factor analysis (CFA). Assessment of a variety of goodness of fit indices offers convergent evidence that the estimated model is statistically a good fit for the data, and offers statistical support for the soundness of the Motivational Consumption Model43: the overall chi-square test of model fit is significant, χ2 (7) = 44.79, p < .001, which is expected given the large sample size of the survey. The root mean square error of approximation (RMSEA) is .07. The comparative fit index (CFI) is .98. The Tucker– Lewis fit index (TLI) is .97. The standardized root mean square residual (SRMR) is .03. Moreover, evaluation of the modification indices revealed no sizable, theoretically meaningful points of ill fit in the estimated model. While the proposed model is driven by the integration of two existing theories, and the abovementioned goodness of fit indices also contribute to its statistical soundness, the authors also tested the fit of all other alternative models in the attempt to rule out alternative explanations,44 and none of the alternative models yielded satisfactory goodness of fit indices, further suggesting that the proposed model is not only theoretically sound, but also empirically robust.

Direct Effects Figure 2 presents the parameter estimates for the coefficients of interest, and standardized coefficients are reported. All of the direct effect hypotheses (H1 through H7) proposed in this study are supported by the data. The data suggest that having favorable attitudes toward the news and motivation to consume news lead to more frequent consumption of news on politics, public affairs, and current events, with news attitudes being the stronger influence of the two. In turn, news beliefs and news motivations positively influence the formation of favorable news attitudes, with news beliefs being the stronger predictor of the two. Moreover, positive news beliefs also lead to heightened news motivations. Furthermore, age and education positively influence news beliefs. Specifically, for every one standard deviation increase in age, news beliefs will increase in the positive direction by .12 standard deviation, p < .001, while holding education constant (H1). For every one standard deviation increase in education, news beliefs will increase in the positive direction by 0.16 standard deviation, p < .001, while holding age constant (H2). For every one standard deviation increase in news beliefs, news attitudes will become more favorable by 0.65 standard deviation, p <

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 715

Figure 2. SEM analyses of direct paths in the Motivational Consumption Model. Note. Maximum likelihood estimation of simultaneous multiple regression analyses were performed to test all direct paths in the Motivational Consumption Theory. Standardized coefficients are reported in Figure 2. χ2(7) = 44.79, p <.001. RMSEA = .07. CFI = .98. TLI = .97. SRMR = .03. SEM = structural equation modeling; RMSEA = root mean square error of approximation; CFI = comparative fit index; TLI = Tucker–Lewis fit index; SRMR = standardized root mean square residual. ***p < .01.

.001, while holding news motivations constant (H3). For every one standard deviation increase in news beliefs, news motivations will increase by 0.76 standard deviation, p < .001 (H4). For every one standard deviation increase in news motivations, news attitudes will become more favorable by 0.20 standard deviation, p < .001, while holding news beliefs constant (H5). For every one standard deviation increase in favorable news attitudes, news consumption will increase by 0.34 standard deviation, p < .001, while holding news motivations constant (H6). For every one standard deviation increase in news motivations, news consumption will increase by 0.18 standard deviation, p < .001, while holding news attitudes constant (H7). The variables included in Figure 2 account for 65.9% of variance in news attitudes, 57.4% of variance in news motivations, 23.8% of variance in news consumption, and 4.2% variance in news beliefs.

Indirect Effects As Table 1 suggests, all indirect effects tested in this study are statistically significant, which offers further support for the workings of the Motivational Consumption Model.45 Specifically, age positively affects news consumption through news beliefs and news motivations, b = .02, p < .01, through news beliefs and news attitudes, b = .03, p < .01, as well as through news beliefs, news motives, and news attitudes, b = .01, p < .01 (H8). Whereas H8 examines age’s indirect effects on news consumption, H9 investigates education’s indirect effects on news consumption. In particular, education has a positive effect on news consumption through news beliefs and news motivations, b = .02, p < .01, through news beliefs and news attitudes, b = .04, p < .01, as well as through news beliefs, news motives, and news attitudes, b = .01, p < .01. On the other

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 716 Journalism & Mass Communication Quarterly 91(4)

Table 1. Mediation Analyses of the Motivational Consumption Model (N = 1,143).

Indirect effects among exogenous and endogenous variables b Age to news beliefs to news motivations to news consumption .02** Age to news beliefs to news attitudes to news consumption .03** Age to news beliefs to news motives to news attitudes to news .01** consumption Education to news beliefs to news motivations to news consumption .02** Education to news beliefs to news attitudes to news consumption .04*** Education to news beliefs to news motivations to news attitudes to news .01*** consumption News beliefs to news motivations to news consumption .14*** News beliefs to news attitudes to news consumption .22** News beliefs to news motivations to news attitudes to news consumption .05***

Note. Cell entries are standardized coefficients from maximum likelihood estimation of simultaneous multiple regression analyses under the structural equation modeling (SEM) framework. **p < .01. ***p < .001. hand, H10 tests all mediation paths between news beliefs and news consumption in the structural model: news beliefs positively affect news consumption through news motivations, b = .14, p < .001, through news attitudes, b = .22, p < .01, as well as through news motivations and news attitudes, b = .05, p < .01.

Discussion Given the importance of having an informed democratic society, news consumption is a crucial theoretical construct that necessitates empirical examination. Whereas the uses and gratifications paradigm offers a rich understanding of different gratifications or motivations associated with media use, one of its enduring criticisms remains its inability to predict behavior and contribute to theoretical advancement.46 On the other hand, the RAM is strong in its ability to predict social behavior, although it has rarely been used outside of health communication. Contributing to our understanding of news consumption, this study offers a novel way of understanding the psychology behind news consumption by integrating the uses and gratifications paradigm and the RAM. The proposed model seeks to explain why age and education positively influence news consumption by proposing a model with three psychological mediators (news beliefs, news motivations, and news attitudes). It is not getting older or obtain- ing one more diploma that leads to people’s consuming more news per se. Instead, this study finds such changes are psychologically mediated by one’s news beliefs, news motivations, and news attitudes. Using SEM, an advanced class of statistical analyses, this study offers statistical support for the theoretical underpinning of the Motivational Consumption Model. The data suggest that (1) age and education lead to positive beliefs about news; (2) the more positive beliefs one has about news, the more one will be motivated to consume news

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 717

and the more favorable attitudes one will have toward news; and (3) the more one is motivated to consume news and the more favorable attitudes one has toward news, the more frequently one will consume news on politics, public affairs, and current events. Overall, the model explains 23.8% of the variance in news consumption, which is significantly higher than many media use studies—those based on uses and gratifications, for example, typically account for less than 10% of the variance.47 Moreover, a meta-analysis on behavior prediction suggests that variances explained in the 20% range are considered a medium-to-larger effect size.48 Granted, this model accounts only for psychological influences on news consumption. To develop a more comprehensive model, future studies are encouraged to examine how other factors influence existing theoretical linkages. Despite its theoretical contribution, this study is not without limitations. There are a few methodological constraints: While the news consumption variable comprises eleven news sources that primarily carry news on politics, public affairs news, and current events, the measure used in this study cannot ensure that the respondents do not use these traditional news outlets for other purposes (i.e., for crossword puzzles or entertainment news). Future studies should explore other ways to better measure consumption of news on politics, public affairs news, and current events. Also, the news consumption variable is measured at the interval level (i.e., never, seldom, sometimes, and often). While this level of measurement is purposefully chosen to better reflect the natural state in which average respondents think about news consumption, which has been shown to minimize reporting errors,49 it is limited in its measurement precision (compared to time spent in minutes or hours). Future studies should test other media use measures. Moreover, research suggests that average respondents tend to over- report their news consumption level.50 Using self-reports for news consumption, this study is also subject to the halo effect, and its findings should thus be interpreted with this possible threat in mind. This study has taken some measures to minimize this threat, however. For example, rather than using a single measure, this study asked respondents about their news consumption habits across thirty news outlets, including non-traditional news outlets such as late-night talk shows, political satire, and social media sites. This extensive list of news outlets arguably allows for a more accurate measure of average respondents’ news use. Nonetheless, future studies should examine possible halo effects associated with this type of measure. Although this study used an online panel survey that matches favorably to data from the U.S. census51 on four key demographic variables (age, gender, income, and race), panel-based online surveys are ultimately non-probability in nature. As such, findings should be interpreted with more emphasis on the strength of relationships among key variables than on statistical significance. Future studies should use surveys involving probability sampling or experiments. Also, while this study specifically focuses on attitudes, future studies may investigate how education may indirectly influence news consumption through norms and self-efficacy. Future studies should also examine whether the Motivational Consumption Model applies to consumption of other genres of news, such as soft news or niche news (i.e., finance). After all, despite the information value of news content from legacy news media outlets, studies

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 718 Journalism & Mass Communication Quarterly 91(4) have found that the public also acquires important political information from other sources, such as political satire or late-night entertainment talk shows.52 The Motivational Consumption Model does not solve all the problems and questions that come with news consumption. For example, the current model does not incorporate price as a factor because, among the eleven types of news media included in this study, only two (national and local newspapers) have a price tag, and the rest are offered for free (e.g., local TV newscasts, Google News, Yahoo News, etc.). Therefore, one can assume that the paywall factor has limited influence on the use of most news media outlets examined in this study. Nevertheless, price is an important factor influencing media consumption when it comes to fee-based content. In addition, news noteworthiness53—the interaction of taste (preference) with available content—has also been found to influence news use and paying intent. Future studies should expand the model to include these factors when studying news consumption. Beyond theoretical contributions, the Motivational Consumption Model also car- ries practical implications for intervention because it comprises psychological and behavioral factors that can be influenced. At the most fundamental level, while age is contingent on time alone and cannot be changed through interventions, education is amendable in that policy makers can mandate more formative education for the public, and educators can be reminded that their teaching makes a difference in encouraging news consumption. Moreover, in terms of news beliefs, developing education programs that emphasize the importance and empowerment of news can serve as the foundation in promoting favorable attitudes toward the news and heightening news motivations, both of which are significant predictors of news consumption. Overall, through its integration of existing theories, the Motivational Consumption Model provides a way to understand news consumption through a psycho-behavioral lens. While this study is by no means comprehensive, it offers a parsimonious, test- able, and generalizable model with notable explanatory power, paving the way for further theoretical development on news consumption.

Acknowledgments The authors thank Dominic Lasorsa, Yachu Liu, Regina Lawrence, and anonymous reviewers for offering helpful advice on this study.

Authors’ Note An earlier version of the manuscript received the 2012 Patricia Witherspoon Research Award from The Annett Strauss Institute for Civic Life.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 719

Notes 1. Judith Moeller, Claes de Vreese, Frank Esser, and Ruth Kunz, “Pathway to Political Participation: The Influence of Online and Offline News Media on Internal Efficacy and Turnout of First-Time Voters,” American Behavioral Scientist 58 (5, 2014): 689- 700, doi:10.1177/0002764213515220; Paula Poindexter, “Factors Contributing to the Sex Divide in Newspapers and Television News,” in Women, Men, and News: Divided and Disconnected in the News Media Landscape, ed. Paula Poindexter, Sharon Meraz, and Amy Schmitz Weiss (NY: Routledge, 2008), 17-34; Dietram A. Scheufele, James Shanahan, and Sei-Hill Kim, “Who Cares about Local Politics? Media Influences on Local Political Involvement, Issue Awareness, and Attitude Strength,” Journalism & Mass Communication Quarterly 79 (June 2002): 427-44, doi:10.1177/107769900207900211. 2. Scheufele, Shanahan, and Kim, “Who Cares about Local Politics?” 3. Michael X. Delli Carpini and Scott Keeter, What Americans Know about Politics and Why It Matters (New Haven, CT: Yale University Press, 1996); Gianpietro Mazzoleni and Winfried Schulz, “‘Mediatization’ of Politics: A Challenge for Democracy?” Political Communication 16 (3, 1999): 247-61, doi:10.1080/105846099198613. 4. Shanto Iyengar and Donald R. Kinder, News That Matters: Television and American Opinion (Chicago: University of Chicago Press, 1987); Markus Prior, “The Immensely Inflated News Audience: Assessing Bias in Self-Reported News Exposure,” Public Opinion Quarterly 73 (1, 2009): 130-43; Shanto Iyengar, Is Anyone Responsible? How Television Frames Political Issues (Chicago: University of Chicago Press, 1991); Maxwell McCombs, Setting the Agenda: The News Media and Public Opinion (Oxford, UK: Blackwell, 2004). 5. Excluding cross-platform estimates. 6. Andrew Beaujon, “Pew: 17% of Americans Get No News Daily,” Poynter, April 27, 2012, http://www.poynter.org/latest-news/mediawire/171941/pew-17-of-americans-get-no- news-daily/; Pew Research Center for the People & the Press, “Americans Spending More Time Following the News,” September 12, 2010, http://www.people-press.org/2010/09/12/ americans-spending-more-time-following-the-news/. 7. Pew Research Center for the People & the Press, “Americans Spending More Time Following the News.” 8. Judee K. Burgoon and Michael Burgoon, “Predictors of Newspaper Readership,” Journalism Quarterly 57 (winter 1980): 589-96; Klaus Schoenbach, Edmund Lauf, Jack M. McLeod, and Dietram A. Scheufele, “Distinction and Integration: Sociodemographic Determinants of Newspaper Reading in the USA and Germany, 1974-96,” European Journal of Communication 14 (2, 1999): 225-39; Robert L. Stevenson, “The Frequency of Newspaper Readership,” ANPA News Research Report 7 (1977): 1-8; Bruce H. Westley and Werner J. Severin, “A Profile of the Daily Newspaper Non-Reader,” Journalism Quarterly 41 (4, 1964): 45-50, 156. 9. Scott L. Althaus and David Tewksbury, “Patterns of Internet and Traditional News Media Use in a Networked Community,” Political Communication 17 (1, 2000): 21-45; Deborah S. Chung, “Interactive Features of Online Newspapers: Identifying Patterns and Predicting Use of Engaged Readers,” Journal of Computer-Mediated Communication 13 (3, 2008): 658-79; Hsiang Iris Chyi, Menchieh Jacie Yang, Seth C. Lewis, and Nan Zheng, “Use of and Satisfaction with Newspaper Sites in the Local Market: Exploring Differences between Hybrid and Online-Only Users,” Journalism & Mass Communication Quarterly 87 (1, 2010): 62-83; Barbara K. Kaye and Thomas J. Johnson, “Online and in the Know:

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 720 Journalism & Mass Communication Quarterly 91(4)

Uses and Gratifications of the Web for Political Information,” Journal of Broadcasting & Electronic Media 46 (1, 2002): 54-71, doi:10.1207/s15506878jobem4601_4; Daniel Riffe, Stephen Lacy, and Miron Varouhakis, “Media System Dependency Theory and Using the Internet for In-Depth, Specialized Information,” Web Journal of Mass Communication Research 11 (2008): 1-14; Robert LaRose and Matthew S. Eastin, “A Social Cognitive Theory of Internet Uses and Gratifications: Toward a New Model of Media Attendance,” Journal of Broadcasting & Electronic Media 48 (3, 2004): 358-77, doi:10.1207/ s15506878jobem4803_2. 10. Oscar Westlund, “From Mobile Phone to Mobile Device: News Consumption on the Go,” Canadian Journal of Communication 33 (3, 2008): 443-63; Ran Wei, Ven-hwei Lo, Xiaoge Xu, Yi-Ning Katherine Chen, and Guoliang Zhang, “Predicting Mobile News Use among College Students: The Role of Press Freedom in Four Asian Cities,” New Media & Society 16 (6, 2013): 637-54, doi:10.1177/1461444813487963; Hsiang Iris Chyi and Monica Chadha, “News on New Devices: Is Multi-platform News Consumption a Reality?,” Journalism Practice 6 (4, 2012): 431-49; Sylvia Chan-Olmsted, Hyejoon Rim, and Amy Zerba, “Mobile News Adoption among Young Adults: Examining the Roles of Perceptions, News Consumption, and Media Usage,” Journalism & Mass Communication Quarterly 90 (1, 2013): 126-47, doi:10.1177/1077699012468742. 11. Christopher M. Raine, “Uses and Gratifications of Facebook for Political Information” (Master’s thesis, University of Kansas, Lawrence, 2008). 12. While other demographic variables also affect news beliefs, this theory explores only two of these variables that are the most central to the creation and maintenance of beliefs toward news on politics, public affairs, and current events; news beliefs are the founda- tional building block in the Motivational Consumption Model that indirectly influence news consumption through news attitudes and news motivations. 13. Pew Research Center for the People & the Press, “Americans Spending More Time Following the News”; Paula M. Poindexter and Maxwell E. McCombs, “Revisiting the Civic Duty to Keep Informed in the New Media Environment,” Journalism & Mass Communication Quarterly 78 (1, 2001): 113-26. 14. Arvind Diddi and Robert LaRose, “Getting Hooked on News: Uses and Gratifications and the Formation of News Habits among College Students in an Internet Environment,” Journal of Broadcasting & Electronic Media 50 (2, 2006): 193-210, doi:10.1207/ s15506878jobem5002_2; Alan M. Rubin and Elizabeth M. Perse, “Audience Activity and Television News Gratifications,” Communication Research 14 (1, 1987): 58; Elihu Katz, Hadassah Haas, and Michael Gurevitch, “On the Use of the Mass Media for Important Things,” American Sociological Review 38 (2, 1973): 164-81; Maxwell McCombs and Paula Poindexter, “The Duty to Keep Informed: News Exposure and Civic Obligation,” Journal of Communication 33 (2, 1983): 88-96, doi:10.1111/j.1460-2466.1983.tb02391.x; Denis McQuail, Jay G. Blumler, and J. Brown, “The Television Audience: A Revised Perspective,” in Sociology of Mass Communication, ed. Denis McQuail (Middlesex, UK: Penguin, 1972), 135-65. 15. Jay D. Hmielowski, R. Lance Holbert, and Jayeon Lee, “Predicting the Consumption of Political TV Satire: Affinity for Political Humor, The Daily Show, and The Colbert Report,” Communication Monographs 7 (1, 2011): 96-114; Katz, Haas, and Gurevitch, “On the Use of the Mass Media for Important Things”; Carolyn A. Lin, Michael B. Salwen, and Rasha A. Abdulla, “Uses and Gratifications of Online and Offline News: New Wine in an Old Bottle,” in Online News and the Public, ed. Michael B. Salwen, Bruce Garrison, and Paul

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 721

D. Driscoll (Mahwah, NJ: Lawrence Erlbaum, 2005), 221-36; Alan M. Rubin, “Ritualized and Instrumental Television Viewing,” Journal of Communication 34 (3, 1984): 67-77. 16. For example, R. Kelly Garrett, “Echo Chambers Online? Politically Motivated Selective Exposure among Internet News Users,” Journal of Computer-Mediated Communication 14 (2, 2009): 265-85; Shanto Iyengar and Kyu S. Hahn, “Red Media, Blue Media: Evidence of Ideological Selectivity in Media Use,” Journal of Communication 59 (1, 2009): 19-39; Thomas J. Johnson, Shannon L. Bichard, and Weiwu Zhang, “Communication Communities or ‘Cyberghettos?’ A Path Analysis Model Examining Factors That Explain Selective Exposure to Blogs,” Journal of Computer-Mediated Communication 15 (1, 2009): 60-82; Barbara K. Kaye, “Web Site Story: An Exploratory Study of Blog Use Motivations,” in Blogging, Citizenship, and the Future of Media, ed. Mark Tremayne (NY: Routledge, 2007), 127-48; Silvia Knobloch-Westerwick and Jingbo Meng, “Looking the Other Way: Selective Exposure to Attitude-Consistent and Counterattitudinal Political Information,” Communication Research 36 (3, 2009): 426-48, doi:10.1177/0093650209333030; Natalie Jomini Stroud, “Media Use and Political Predispositions: Revisiting the Concept of Selective Exposure,” Political Behavior 30 (3, 2008): 341-66; Natalie Jomini Stroud, Niche News: The Politics of News Choice (NY: Oxford University Press, 2011). 17. Philip Palmgreen and J. D. Rayburn, “Uses and Gratifications and Exposure to Public Television: A Discrepancy Approach,” Communication Research 6 (2, 1979): 155-79, doi:10.1177/009365027900600203; Angela M. Lee, “The Future of News? Examining Ritualistic News Consumption in Traditional and New Media” (paper presented at the 95th Annual Convention of the National Communication Association, Chicago, IL, 2009), http://citation.allacademic.com/meta/p_mla_apa_research_citation/3/2/0/5/3/p320534_ index.html?phpsessid=2126d2ot994s8c2fd4rt34ggb2. 18. Angela M. Lee, “News Audiences Revisited: Theorizing the Link between Audience Motivations and News Consumption,” Journal of Broadcasting & Electronic Media 57 (3, 2013): 300-17, doi:10.1080/08838151.2013.816712. 19. Pamela J. Shoemaker, “Hardwired for News: Using Biological and Cultural Evolution to Explain the Surveillance Function,” Journal of Communication 46 (3, 1996): 32-47, doi:10.1111/j.1460-2466.1996.tb01487.x; Matthias R. Hastall, “Informational Utility as Determinant of Media Choices,” in Media Choice: A Theoretical and Empirical Overview, ed. Tilo Hartmann (NY: Routledge, 2009), 149-66. 20. Amy Bleakley, Martin Fishbein, Michael Hennessy, and Amy Jordan, “How Sources of Sexual Information Relate to Adolescents’ Beliefs about Sex,” American Journal of Health Behavior 33 (1, 2009): 37; Icek Ajzen, “Martin Fishbein’s Legacy: The Reasoned Action Approach,” The ANNALS of the American Academy of Political and Social Science 640 (1, 2012): 11-27; Dolores Albarracin, Blair T. Johnson, Martin Fishbein, and Paige A. Muellerleile, “Theories of Reasoned Action and Planned Behavior as Models of Condom Use: A Meta-analysis,” Psychological Bulletin 127 (1, 2001): 142; Christopher J. Armitage and Mark Conner, “Efficacy of the Theory of Planned Behaviour: A Meta-analytic Review,” British Journal of Social Psychology 40 (4, 2001): 471-99; Amy Bleakley and Michael Hennessy, “The Quantitative Analysis of Reasoned Action Theory,” The ANNALS of the American Academy of Political and Social Science 640 (1, 2012): 28-41. 21. Martin Fishbein and Icek Ajzen, Predicting and Changing Behavior (NY: Taylor & Francis, 2010); Rebecca L. Collins, Marc N. Elliott, Sandra H. Berry, David E. Kanouse, Dale Kunkel, Sarah B. Hunter, and Angela Miu, “Watching Sex on Television Predicts Adolescent Initiation of Sexual Behavior,” Pediatrics 114 (3, 2004): 280-89; S. Liliana

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 722 Journalism & Mass Communication Quarterly 91(4)

Escobar-Chaves, Susan R. Tortolero, Christine M. Markham, Barbara J. Low, Patricia Eitel, and Patricia Thickstun, “Impact of the Media on Adolescent Sexual Attitudes and Behaviors,” Pediatrics 116 (suppl. 1, 2005): 303-26; Bleakley et al., “How Sources of Sexual Information Relate to Adolescents’ Beliefs about Sex.” 22. Martin Fishbein and Icek Ajzen, Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research (Reading: MA: Addison-Wesley, 1975), http://trid.trb.org/view. aspx?id=1150648; Ajzen, “Martin Fishbein’s Legacy”; Fishbein and Ajzen, Predicting and Changing Behavior. 23. Fishbein and Ajzen, Predicting and Changing Behavior. 24. For example, Paschal Sheeran and Sheina Orbell, “Do Intentions Predict Condom Use? Metaanalysis and Examination of Six Moderator Variables,” British Journal of Social Psychology 37 (2, 1998): 231-50; Gaston Godin and Gerjo Kok, “The Theory of Planned Behavior: A Review of Its Applications to Health-Related Behaviors,” American Journal of Health Promotion 11 (2, 1996): 87-98; Armitage and Conner, “Efficacy of the Theory of Planned Behaviour”; Albarracin et al., “Theories of Reasoned Action and Planned Behavior as Models of Condom Use.” 25. Bleakley and Hennessy, “The Quantitative Analysis of Reasoned Action Theory.” 26. Ajzen, “Martin Fishbein’s Legacy.” 27. McCombs and Poindexter, “The Duty to Keep Informed.” 28. Vincent Guilamo-Ramos, James Jaccard, Patricia Dittus, Bernardo Gonzalez, and Alida Bouris, “A Conceptual Framework for the Analysis of Risk and Problem Behaviors: The Case of Adolescent Sexual Behavior,” Social Work Research 32 (1, 2008): 29-45; James Jaccard, David Brinberg, and Lee J. Ackerman, “Assessing Attribute Importance: A Comparison of Six Methods,” Journal of Consumer Research 12 (4, 1986): 463-68. 29. Jane Sasseen, Kenny Olmstead, and Amy Mitchell, “Digital: By the Numbers,” The State of the News Media 2013, 2013, http://stateofthemedia.org/2013/ digital-as-mobile-grows-rapidly-the-pressures-on-news-intensify/digital-by-the-numbers/. 30. Fishbein and Ajzen, Predicting and Changing Behavior. 31. Martin Fishbein, “A Behavior Theory Approach to the Relations between Beliefs about an Object and the Attitude toward the Object,” in Readings in Attitude Theory and Measurement, ed. Martin Fishbein (NY: Wiley, 1967), 389-99; Fishbein and Ajzen, Predicting and Changing Behavior; Ajzen, “Martin Fishbein’s Legacy.” 32. Poindexter, “Factors Contributing to the Sex Divide in Newspapers and Television News”; Philip Palmgreen, Lawrence A. Wenner, and Karl E. Rosengren, “Uses and Gratifications Research: The Past Ten Years,” in Media Gratifications Research: Current Perspectives, ed. Karl E. Rosengren, Lawrence A. Wenner, and Philip Palmgreen (Beverly Hills, CA: SAGE, 1985), 11-37, http://www.citeulike.org/group/14507/article/9360868. 33. Alan M. Rubin and Mary M. Step, “Impact of Motivation, Attraction, and Parasocial Interaction on Talk Radio Listening,” Journal of Broadcasting & Electronic Media 44 (4, 2000): 635-54; Thomas E. Ruggiero, “Uses and Gratifications Theory in the 21st Century,” Mass Communication and Society 3 (1, 2000): 3-37; Natalie Kink and Thomas Hess, “Search Engines as Substitutes for Traditional Information Sources? An Investigation of Media Choice,” The Information Society 24 (1, 2008): 18-29, doi:10.1080/01972240701771630; William J. McGuire, “Psychological Motives and Communication Gratification,” in The Uses of Mass Communications: Current Perspectives on Gratifications Research, ed. Jay G. Blumler and Elihu Katz (Beverly Hills, CA: SAGE, 1974), 167-96. 34. Elihu Katz, “Mass Communication Research and the Study of Popular Culture,” Studies in Public Communication 2 (1, 1959): 1-6; Elizabeth M. Perse and Debra Greenberg

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Lee and Chyi 723

Dunn, “The Utility of Home Computers and Media Use: Implications of Multimedia and Connectivity,” Journal of Broadcasting & Electronic Media 42 (4, 1998): 435-56; Denis McQuail, Mass Communication Theory: An Introduction, 3rd ed. (London, England: SAGE, 1994). 35. Palmgreen and Rayburn, “Uses and Gratifications and Exposure to Public Television.” 36. Hmielowski, Holbert, and Lee, “Predicting the Consumption of Political TV Satire”; Rubin, “Ritualized and Instrumental Television Viewing”; Katz, Haas, and Gurevitch, “On the Use of the Mass Media for Important Things”; Lin, Salwen, and Abdulla, “Uses and Gratifications of Online and Offline News.” 37. McQuail, Blumler, and Brown, “The Television Audience”; McCombs and Poindexter, “The Duty to Keep Informed”; Shoemaker, “Hardwired for News”; Rubin, “Ritualized and Instrumental Television Viewing”; Rubin and Perse, “Audience Activity and Television News Gratifications”; Diddi and LaRose, “Getting Hooked on News.” 38. Hsiang Iris Chyi and Angela M. Lee, “Online News Consumption: A Structural Model Linking Preference, Use, and Paying Intent,” Digital Journalism 1 (2, 2013): 194-211. 39. Office of Survey Research is a member of the Association of Academic Survey Research Organizations (AASRO) and the American Association for Public Opinion Research (AAPOR). Office of Survey Research has offered expert survey research services to academics, professionals, and the government since 1986. 40. U.S. Census Bureau, “American Fact Finder,” 2012, http://factfinder2.census.gov. 41. Completion rate is number of survey respondents who provided usable responses divided by the total number of respondents who clicked into the survey. 42. Cronbach’s alpha of .7 or greater is considered satisfactory; see Jack Cohen, Statistical Power Analysis for the Behavioral Sciences (Hillsdale, NJ: Lawrence Erlbaum, 1988). 43. Conventional structural equation modeling (SEM) standards, contingent on sample size, suggest the following to be indication of good model fits: (1) non-significant chi-square test, (2) root mean square error of approximation (RMSEA) less than .08, (3) comparative fit index (CFI) equal or greater than .9, (4) Tucker–Lewis fit index (TLI) equal to or greater than .9, (5) standardized root mean square residual (SRMR) less than .05. 44. For example, whereas in the current model, news beliefs is the predictor of news attitudes and motivations, the author tested alternative models that rotated the causal order of news beliefs, news attitudes, and news motivation, and none of those alternative models yields satisfactory goodness of fit tests. 45. For visual simplicity, indirect effects are excluded from the path model. 46. McQuail, Mass Communication Theory: An Introduction; Cheryl Bracken and Matthew Lombard, “Uses and Gratifications: A Classic Methodology Revisited,” New Jersey Journal of Communication 9 (1, 2001): 103-16; Oscar Peters, Matthias Rickes, Sven Jöckel, Christian von Criegern, and Alexander van Deursen, “Explaining and Analyzing Audiences: A Social Cognitive Approach to Selectivity and Media Use,” Communications 31 (3, 2006): 279-308; Daniel G. McDonald, “Media Orientation and Television News Viewing,” Journalism & Mass Communication Quarterly 67 (1, 1990): 11-20, doi:10.1177/107769909006700103. 47. LaRose and Eastin, “A Social Cognitive Theory of Internet Uses and Gratifications.” 48. Armitage and Conner, “Efficacy of the Theory of Planned Behaviour.” 49. Pamela J. Shoemaker, James W. Tankard, and Dominic L. Lasorsa, How to Build Social Science Theories (Thousand Oaks, CA: SAGE, 2004). 50. Prior, “The Immensely Inflated News Audience.”

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 724 Journalism & Mass Communication Quarterly 91(4)

51. U.S. Census Bureau, “American Fact Finder.” 52. Barry A. Hollander, “Late-Night Learning: Do Entertainment Programs Increase Political Campaign Knowledge for Young Viewers?” Journal of Broadcasting & Electronic Media 49 (4, 2005): 402-15; Lauren Feldman, “The News about Comedy Young Audiences, The Daily Show, and Evolving Notions of Journalism,” Journalism 8 (4, 2007): 406-27. 53. Angela M. Lee and Hsiang Iris Chyi, “When Newsworthy Is Not Noteworthy: Examining the Value of News from the Audience’s Perspective,” Journalism Studies (2013), doi:10.1080/1461670X.2013.841369.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 JMQXXX10.1177/1077699014550093Journalism & Mass Communication QuarterlySuran et al. 550093research-article2014

Audience Motivations and Responses

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 725–739 Topical Punch: Health © 2014 AEJMC Reprints and permissions: Topics as Drivers of sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550093 Idiosyncratic Reader jmcq.sagepub.com Responses to Online News

Melissa Suran1, Avery E. Holton2, and Renita Coleman1

Abstract Although scholars explore how news stories’ framing elements may affect reader responses, they have yet to examine how the topics of health-related articles affect those responses. By content analyzing three US newspapers’ online health content and reader comments, this study finds that certain health topics are idiosyncratic with reader responses. Readers reacted to personal health and obesity news with more episodic and gain-framed comments but relied more on loss frames when discussing chronic health issues. Readers also used more thematic frames in comments about mental illness. Health coverage related to politics and the government was associated with fewer episodic comments.

Keywords episodic and thematic framing, gain and loss framing, health frames, health journalism, health news, online comments, news consumers

Discourse surrounding health concerns has remained among the top news issues in the United States over the last several years, particularly online where traffic to news sites continues to rise.1 These outlets serve as important conduits of health content and provide a certain level of public engagement, often offering spaces where readers can comment on reports and discuss related issues. Allowing readers to comment on online

1The University of Texas at Austin, Austin, TX, USA 2The University of Utah, Salt Lake City, UT, USA Corresponding Author: Melissa Suran, The University of Texas at Austin, School of Journalism, 300 W. Dean Keeton St., Austin, TX 78712-1073, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Downloaded from http://www.elearnica.ir 726 Journalism & Mass Communication Quarterly 91(4) news coverage opens up public deliberation, exposes people to novel ideas, and invites a variety of perspectives and opinions that may not be included in the news coverage itself.2 Several recent studies have examined the association between the ways health news articles are framed and the ways readers respond, finding that readers may pick up on particular frames but may not promote them in online discourse.3 This is particularly important in light of research that shows news consumers who read online content and the comments that follow may align their beliefs and opinions more with reader responses and less with actual news coverage.4 Even with balanced reporting, reader comments have the potential to skew news consumers’ perceptions. While public health officials and scholars have urged journalists for years to improve their health news coverage, especially in ways that can help people connect with health issues on societal rather than individual levels,5 reader comments have only recently come under fire. Indeed, journalists have taken steps to enhance their health coverage by including more societal, or thematic, frames in their reporting.6 Readers, however, have not entirely embraced the change, continuing to approach some health issues as individual problems in spite of the thematic frames offered by journalists.7 This suggests that certain topics, rather than framing devices, may drive reader responses. For example, the topic of obesity may engender episodic thinking no matter how much thematic framing journalists employ. Thus, this study questions how the topic of health news content, rather than the frames, may guide reader responses. If it is the case that health topics influence the ways in which readers react, then improving health news coverage may need to extend beyond simply shifting framing techniques. Using a content analysis of online health articles and reader responses from three major U.S. news outlets, this study explores the association between the topic of health content and the frames with which readers choose to respond. The findings signify the importance of particular health topics in promoting certain forms of reader responses and can help health communication scholars and journalists design potentially more effective approaches to health news coverage by anticipating certain forms of reader reactions.

Literature Review Over the last several decades, the Internet has opened up new forms of news media that help to inform the public while also promoting a certain level of dialogue between news creators and consumers.8 News creators have attempted to harness this new form of news interaction, navigating into social media spaces and developing forums for public conversations. While these spaces may not always provide balanced engagement between members of the news media and their audiences,9 they do allow individuals an opportunity to extend conversations and offer reactions. News organizations may continue to graft traditional journalism routines into digital spaces, keeping the public out of certain aspects of the news production process, but they do so at the risk of losing touch with what the public values as news, as well as public engagement with the news.10 At least one study shows that reader commentary following news articles, particularly those about science and health, can have a more profound effect on readers than the content of the journalistically-produced product.11

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Suran et al. 727

Several other studies have approached this issue from a health communication angle, exploring how framing techniques used by health journalists can alter the ways in which readers respond to those articles.12 These studies indicate that journalists have amended, at least in part, the way they cover certain health issues through their framing choices. However, these adjustments have not always resulted in congruent changes in reader responses. Instead, the findings from these studies call into question the role of specific health topics (e.g., cancer, HIV/AIDS, mental health, sexual health) in reader comments. More specifically, they raise the question of whether health topics may influence reader responses as much as or more than news frames.

Health Frames and Reader Responses Communication scholars have pointed out the importance of reader responses to the news, in particular, the function of reader comments to online news coverage that can potentially reach and influence anyone on the Internet.13 A growing number of Internet users continue to turn to the media for health information, making the content they encounter—including reader comments—of particular importance. This study is not the first to call into question the role of health topics in relation to perceptions of or reactions to key health issues, but it is the first to date to empirically explore the association between health topics in news coverage and the ways in which readers respond to those issues. Many studies of health news have worked within the theoretical construct of framing, which can be defined as the structural devices or selection of information used by journalists to explain issues or events.14 Of particular interest, health scholars have given much focus to two dichotomous framing mechanisms often used by health reporters: thematic/episodic framing and gain/loss framing.15 Thematic frames connect issues and events with society and emphasize shared responsibility for solving problems, while episodic frames emphasize individual roles in society, often attribut- ing responsibility to a single individual.16 It is also important to note that health experts have a strong partiality toward thematic framing, often arguing that journalists do not incorporate enough thematic framing in their stories.17 The thesis of this argument is that the inclusion of thematic frames can cue the public to think about issues on a societal rather than individual level, helping overcome the human tendency to blame individuals, known as the fundamental attribution error. These two different types of frames accomplish different purposes. For example, episodic framing puts human faces on stories and attracts attention. Thematic frames provide information about what society can do; without them, readers acquire an incomplete understanding of health issues. A primary function of journalists is to balance their reporting. Thus, the objective of balanced framing should not be to use only episodic framing or only thematic framing, but rather to achieve a balance of both. Still, a large contingent of health scholars has suggested that the news media can help improve public health knowledge by emphasizing thematic frames in order to help individuals understand societal causes, since it is currently lacking in health coverage.18 Indeed, research demonstrates that the use of such frames encourages people to assign

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 728 Journalism & Mass Communication Quarterly 91(4) responsibility for health issues, such as obesity, to society rather than to individuals.19 Other research has cautioned against such optimism, finding that thematic frames in online health news may not be consistent with reader responses.20 This is congruent with more general research that shows comments from online news consumers frequently chastise others rather than respond with societal information.21 Health communication scholars have observed that even with balanced news coverage, the public is quick to assume that sick individuals are responsible for their health problems.22 Certain issues (e.g., obesity, diabetes) tend to trigger comments that are less thematic and more episodic.23 Gain framing and loss framing, which have been shown to influence perceptions of and reactions to health communication and issues as well as health behaviors,24 represent another dimension of framing. Gain frames promote the positive outcomes of specific actions, such as cancer screenings or childhood vaccinations, while loss frames focus on the negative outcomes of not partaking in those actions. While gain frames tend to be persuasive for low-risk actions such as taking vitamins or having blood drawn, loss frames are more effective for high-risk actions such as surgery.25 When it comes to the ways readers respond to the incorporation of these frames into health news, however, the results are mixed.26 While the frames health journalists select are important and may affect those who read them, they may not necessarily be the guiding forces behind the ways readers respond. Thus, it is important to understand what other factors influence readers’ understandings of health issues.

Health Topics Health communication researchers are often very topic specific, usually focusing on a single issue, such as autism,27 obesity,28 or HIV/AIDS.29 When studies are more inclu- sive, they sometimes create an excessive number of topics, oftentimes incorporating similar issues, such as nutrition and obesity,30 which can be collapsed into one category, or hepatitis and scoliosis,31 which can be subsumed under the broader category of chronic health issues that include multiple, related subtopics. In their study of two decades’ worth of health coverage in the U.S. media, Manganello and Blake32 found eighteen health categories: (1) violence, (2) tobacco, (3) alcohol, (4) illegal drugs, (5) prescription drugs, (6) obesity and nutrition, (7) body image and eating disorders, (8) sex, (9) injury, (10) AIDS, (11) cancer, (12) aging, (13) death and disability, (14) mental health, (15) health providers and organizations, (16) genetics, (17) women’s health, and (18) an “other” category. They also noted that these categorizations could be logi- cally collapsed, could evolve over time, and should be further explored by researchers. Comparable studies33 have presented similar categorizations, often collapsing some health topics together such as obesity and personal health, women’s health and gender- specific and/or sexual health, and chronic illnesses. These and other studies that have incorporated multiple health topics have found that differences in public responses can be attributed to the topics. For example, Hatley Major34 analyzed attributions of responsibility for obesity and lung cancer in a group of 229 U.S. adults who read news articles that were framed as either thematic or

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Suran et al. 729

episodic and gain or loss. Episodic and loss frames were positively associated with individual attributions of responsibility for obesity, but not for lung cancer, indicating the possibility of preconceived notions of responsibility for certain topics. Another study of interventions for negative behaviors, such as smoking, poor diet, poor exercise, and high alcohol consumption, found that different intervention methods worked best with different health issues, again indicating that varying outcomes—communication responses included—were likely to result from different topics.35 These studies raise questions about the association between health topics in news coverage and reader responses—an association that cannot be generalized from scholarship examining a single issue or event. As public health officials continue urging journalists to adjust the frames they employ to improve responses and behaviors in society, these same officials and journalists should consider that the topics of discourse may guide how individuals respond more so than how journalists frame them. This study represents a first step in determining whether specific health topics drive the ways in which readers frame their responses, asking:

RQ1: What association, if any, is there between the topic of health articles and thematic and episodic frames in reader comments? RQ2: What association, if any, is there between the topic of health articles and gain and loss frames in reader comments?

Method In order to analyze these questions, this study used a content analysis to examine the possible associations between topics in online health news and the frames readers used in their responses via comments. Scholars analyzing similar research questions have indicated that such analyses should include news organizations that facilitate and publish reader feedback alongside health-related content.36 In order to develop an appropriate sample for this study, a list of non-localized (i.e., not geographically-specific) news media websites offering health-related content was compiled using HealthNewsReview.org,37 a nonprofit organization that rates health articles on the websites of twenty-one major media outlets for accuracy, balance, and completeness. Given that there is no comprehensive list of health websites, the authors’ extensive search of the Internet found this site to be the most comprehensive available. From this initial list, websites offering readers the opportunity to comment directly below stories were selected. Websites that offered reader feedback only through Facebook, Twitter, or other social network sites were excluded, along with those that hid comments, did not allow comments, or did not have accessible or significant commenting (e.g., Health.com, The New York Times). In addition, websites that relied heavily on outside content were excluded. For example, CNN.com consistently covers health issues and allows reader comments, but the majority of health articles on the site come second-hand from health sources. Based on these exclusions, three major U.S. news outlets with an online presence were used: The Washington Post, The Huffington Post, and USA Today.

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 730 Journalism & Mass Communication Quarterly 91(4)

Sample Constructed week sampling38 between one and three constructed weeks produces more efficient samples than simple random and consecutive day sampling of newspapers. More recent research shows when sampling health news articles, six constructed weeks may more accurately estimate one- and five-year populations.39 Thus, this study developed six constructed weeks using forty-two randomly selected dates for the July 1, 2010, to July 1, 2011, period. A Google search using the website and/or news outlet name (e.g., The Huffington Post, huffingtonpost.com), the keyword “health,” and the randomly selected dates was used to collect an initial sample. At the time of sampling, articles that did not focus on at least one health issue were excluded from the sample (e.g., vehicle safety, recipes for holiday parties). Letters to the editor, opinion and editorial columns, articles without text (e.g., The Huffington Post’s ‘How To’ health videos), and articles from secondary sources were removed. The resulting sample included 129 articles with an average of 137 reader comments each. Given the volume of comments for some articles and noting that many articles did not receive more than ten comments, the first ten original comments40 of each article were coded in chronological order (i.e., oldest comments first). If there were fewer than ten comments, then all comments were coded. This procedure resulted in 966 total reader comments for the sample. The unit of analysis for comments was the sentence.

Measures Topics. Media scholars have yet to develop a single, agreed-upon list of health categories, instead developing their own categories arising from specific data. For example, Smith41 employed thirteen categories, Manganello and Blake42 had eighteen, and Berry, Wharf-Higgins, and Naylor utilized fourteen.43 There was some overlap—all three had categories for cancer and HIV/AIDS, for example. But there also were notable differences. Manganello and Blake, along with Smith, used a “women’s health” category, but Berry, Wharf-Higgins, and Naylor did not. For this study, we relied on the established categories as much as possible, but created our own categories when topics in our dataset did not fit into previous categorizations. For example, a new category of politics and government arose because of Obamacare, which was frequently in the news during the time period. We also collapsed other scholars’ separate categories of heart disease, diabetes, scoliosis, and other chronic conditions into one “chronic” category, and collapsed their separate categories of SARS (severe acute respiratory syndrome), West Nile Virus, injury, staph infections, flu, tuberculosis, and similar acute illnesses into one “acute” category. Our ten categories were (1) cancer, (2) drugs and alcohol (includes smoking and tobacco), (3) personal health and obesity (e.g., diet and nutrition, personal wellness), (4) HIV/AIDS, (5) gender-specific health (e.g., women’s health, sexual health), (6) chronic health issues (e.g., cardiovascular disease, diabetes, scoliosis), (7) acute health issues (e.g., flu, common cold, personal injuries), (8) politics/government health issues, (9) mental health, and (10) other (e.g., cellphone

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Suran et al. 731 risks, sewage safety). Categories were coded dichotomously: 1 = dominant presence of category, 0 = no dominant presence of category. While an article could contain one or several of the health topics, only the dominant topic was selected. This was identified using cues within the article, including the headline, the lead paragraph, and the overall amount of attention given to a topic.

Episodic/thematic frames. Following Iyengar’s44 research, comments were coded for the number of sentences that contained episodic or thematic frames. Episodic frames focus on a case study, particular instance, or subject to illustrate an issue. For example, one comment from the sample discussing vitamin intake read, “I don’t take vitamins. My baby was fine. I was fine. I eat right and exercise and that’s better than chemicals going into my body.” Thematic frames emphasize broader roles and issues, such as this sentence: “Other recent research has shown that moms who both breastfeed and co-sleep with their babies get better quality sleep, and more of it, than either moms who formula feed or moms who sleep separately from their babies.” A sentence could fall into one or both categories, and some sentences contained neither episodic nor thematic frames.

Gain/loss frames. A gain frame emphasizes the benefits of a certain activity, such as the positive outcomes of screening behaviors or healthy dieting. One comment in the sample stated, “Regular check ups will keep your heart healthy and keep you off the operating table.” A loss frame does the opposite, focusing on the negative outcomes of avoiding screening behaviors or unhealthy dieting.45 As an example, one of the responses read, “You eat fast food, you get fat, you die.” The number of sentences that contained either gain frames or loss frames was recorded. Again, sentences could have one, several, or no instances of either frame.

Coding Reliability and Analysis The researchers developed the initial coding scheme and trained two independent coders in multiple practice sessions using articles and comments not included in the final sample. Slight adjustments were made to the coding scheme to improve reliability. Acceptable Krippendorf’s alpha scores46 of .81 or higher were achieved for all comment variables during the final training session. The sample was then randomly split between both coders, and reliability was measured for 20% of the coded sample.47 Krippendorf’s alpha scores were .88 for coded health topics in articles and .84 or higher for all coded comment variables. Data were analyzed with point-biserial correlations48 given the comparison between the dichotomous topics variable and the interval frames variables.

Results The majority of the 129 articles included in the sample came from USA Today (42.7%). Articles from The Huffington Post (36.4%) and The Washington Post (20.9%)

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 732 Journalism & Mass Communication Quarterly 91(4)

Table 1. Topics in Online Health News Articles.

Number of Percentage of Article topic articles sample Cancer 4 3.1 Drugs and alcohol 7 5.4 Personal health and obesity 44 34.1 HIV/AIDS 3 2.3 Gender-specific health 14 10.9 Chronic health issues 14 10.9 Acute health issues 8 6.2 Politics/government health issues 23 17.8 Mental health 9 7.0 Other 3 2.3 Total N 129 100

comprised the remainder of the sample. Similarly, more comments came from USA Today (48.1%) than from The Huffington Post (34.8%) and The Washington Post (17.1%). There were approximately 137 comments per article (M = 137.4, SD = 308.23). As indicated in Table 1, four topics dominated health news coverage within the sample: (1) obesity and personal health (34.1%), (2) politics/government health issues (17.8%), (3) gender-specific health (10.9%), and (4) chronic health issues (10.9%). To examine RQ1, the association between the health topics of online articles and thematic and episodic frames in reader responses to those articles, point-biserial correlation analyses revealed several significant associations (see Table 2). Issues of personal health and obesity were associated with significantly more episodically-framed comments (rpbi i = .057, p < .01), politics and government issues were associated with significantly fewer episodic comments (rpbi = −.069, p < .05), and mental health topics were associated with significantly more thematic comments (rpbi = .074, p < .05). To examine RQ2, the association between health topics of online articles and gain and loss frames in reader comments responding to those articles, point-biserial correlations revealed two significant relationships. Articles about personal health and obesity were positively associated with gain frames (rpbi = .172, p < .001), while articles focusing on chronic illnesses were positively associated with more loss frames (rpbi = .102, p < .01).

Discussion Public health professionals have implored journalists and news producers to rethink the ways they frame certain health issues, suggesting that changes toward more thematic and gain-framed stories could help improve public knowledge and connectedness with those issues. Researchers have recently observed a shift toward thematic coverage of

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Suran et al. 733

Table 2. Point-Biserial Correlations between Health Topics in Online Articles and Frames Used in Reader Responses to Those Articles.

Comment frames

Article topics Thematic Episodic Gain Loss Cancer −.027 −.031 −.019 −.024 Drugs and alcohol .017 −.011 −.029 −.032 Personal health and obesity .023 .057** .172*** .003 HIV/AIDS .019 −.045 −.020 −.001 Gender-specific health −.047 −.018 .033 −.006 Chronic health issues −.028 .012 −.033 .102** Acute health issues −.044 −.037 −.027 .001 Politics/government health issues .019 −.069* −.024 −.027 Mental health .074* −.010 −.010 −.029 Other −.047 .263** .062 .018

*p < .05. **p < .01. ***p < .001.

health news but have duly noted that these changes have not necessarily improved the way society responds.49 The findings from this study suggest that those responses may be linked with some topics of health news, indicating that the frames news producers choose to cover may not be the only significant elements driving public reactions. Yet, it should also be noted that despite the statistical significance of the findings here, only one health topic was associated with more thematic comments: mental health. At a time when health scholars and practitioners continue to implore health journalists to frame their stories more thematically—a call that journalists appear to be heeding—readers may not be as responsive to such frames as hoped. This would suggest that more consideration, both scholarly and in practice, should be given to the ways in which readers consider and respond to health topics in news media coverage. In terms of the individual findings, some of the relationships were expected. The correlation between episodic responses and topics of personal health and obesity, which included prevention, screenings, diet, and exercise, is shown consistently in studies. People see diet and exercise, as well as obesity and taking care of oneself, as something individuals have complete control over. They are typically not inclined to recognize the role of health insurance or access to affordable health care (e.g., safe, free places to exercise, availability and cost of healthy foods, high concentration of fast-food restaurants in certain neighborhoods) as contributing to these issues. Despite the work of some journalists who have increasingly included this type of thematic information in their articles, the public overwhelmingly continues to believe that ill individuals are responsible for their conditions. Perhaps only time and decades of consistent, thematic framing will correct this—similar to what has occurred with smoking. In seeking ways to help the public balance its views about personal health and obesity between episodic and thematic lenses, other means of altering attitudes must be explored. Journalists may be doing their part, but more of the same does not appear to

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 734 Journalism & Mass Communication Quarterly 91(4) be moving the responsibility needle sufficiently. Among the constructs that may be worth exploring, involvement, personal relevance, and emotion are shown to be important for many different outcomes. Perhaps this is the case with health topics. Just as personal relevance inoculates one against agenda-setting effects, it may also ingrain attitudes of responsibility toward certain health issues. Emotion is an important construct that acknowledges how humans are not always rational, and this is one aspect of health beliefs that is understudied. Stereotypes may also play a role. Perhaps stories aimed at dispelling various stereotypes would be more effective in altering whom people hold responsible for certain health problems than would continuing to hammer home societal responsibility. Moreover, health scholars should speak with health professionals to determine what they believe works to change attitudes and test those insights in future research. The personal health and obesity stories in this study also were significantly correlated with gain-framed comments, showing that readers were likely to have positive attitudes toward such articles and talk about them in terms of benefits to be had rather than risks. This finding is a notable one and suggests that readers may be decreasing the level of vitriol associated with this topic in exchange for more encouraging words. This is not all that surprising, given the effectiveness of positive persuasion on individual health.50 As Cooper51 noted, positive messages not only help people feel encouraged but also help create a sense of belonging and community. Perhaps when it comes to health, individuals are recognizing the power of positivity and its effectiveness as a bonding agent. While journalists may work toward thematically framing issues such as obesity in the United States, they continue to write personal health and obesity articles in a sort of “news-you-can-use” format, implying an individual, how-to-help- yourself approach.52 While this may prompt more episodic responses, such as those seen in this study, it may also have the unexpected benefit of promoting gainful conversations. People may be more inclined to provide well-meaning comments as well as share personal success stories to encourage other readers. Topics of politics and government were significantly associated with fewer episodic comments, which was not unexpected, given the government’s role in health care is a society-wide issue. Nor was the chronic illness association with more loss frames surprising, given that heart disease, diabetes, and the like are long-term challenges that people must cope with and for which there are no cures. The positive correlation between mental health topics and thematically-framed comments also is encouraging. Perhaps this is one topic that journalists have covered as a societal issue for so long that people are beginning to incorporate that into their discourse. This may also be a positive sign for reducing the stigma associated with mental health issues, an area where the news media play an important role. While popular-culture products, such as movies and television programs, often crystallize mental illness through one event or individual, the news media have the opportunity to connect society to issues and concerns it might otherwise be unaware of.53 This is not to suggest that the news media play a more important role than popular culture in improving the discourse and knowledge surrounding mental illness, but rather to promote the news media’s role as a possible outlet for the reduction of the stigmatization

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Suran et al. 735

of mental illnesses. While mental illnesses have been episodically portrayed as violent, dangerous, and disparaging conditions in the past,54 they may now be cast less episodically and more accurately by the news media—a source of public discourse historically critiqued for its focus on individuals with mental illnesses rather than society’s perception of them.55 While the news media most likely are not the sole source of society’s shift toward thematic thoughts of mental illness, this is an area that deserves, at the very least, more consideration by scholars researching communication and dis- abilities. Research to determine why this shift has occurred could also help inform efforts on other topics that are still mired in individual responsibility.

Conclusion While scholars have begun exploring the impact of online news stories on reader comments, this is the first study to date that analyzes how readers respond to specific health topics. This study focused on reader comments in online outlets, providing a foundation for future studies to investigate other platforms, such as Twitter and Facebook, where people can interact with the news and each other. These studies could help provide a wider understanding of how people converse about health issues in mediated venues, as well as interpersonal environments. Scholars may also code articles for thematic/episodic and gain/loss sentences in order to determine if the framing of the articles reflects the framing of the reader responses. Future research may then empirically assess whether an article “sets the agenda” for reader comments or if the framing of an article instead stirs up explicit counter-framing by commenters. Instead of focusing solely on the framing of health stories, research should explore other important constructs that may be influential in changing the way people think about health—personal relevance, emotion, and stereotyping are a few that have been important in changing attitudes about other issues. When these constructs are combined with framing elements, journalists may find their work yields quicker and more long-lasting attitude changes. Moreover, recent research notes the power of personal narratives within health-specific stories. Some research has suggested, for instance, that even thematically-framed stories—those that focus on how societal conditions play the primary role in health outcomes such as obesity—often include “real-person” sources because journalists believe this mechanism makes stories more interesting. Dorfman, Wallack, and others56 have postulated, however, that the inclusion of these individual-focused anecdotes may draw attention to the individual and lead readers to assign responsibility to, rather than empa- thize with, the individual even when the rest of the article makes the case that the individual is not responsible. Other researchers have noted the importance of personal narratives in explaining complex health issues, noting that positioning individuals within a story may improve comprehension and positive health behaviors.57 However, the same research cautions that many variables may factor into how individuals respond to the ways health information is framed, ranging from their emotional states and their prior knowledge to their perceptions of an issue and their media consumption habits.

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 736 Journalism & Mass Communication Quarterly 91(4)

Although it is important for health journalists to thematically frame their articles on all health topics, something about the topic of coverage remains a powerful force in determining how the reader will feel about the piece. Framing theory, especially the thematic/episodic and gain/loss aspect, which is so important to health communication, should incorporate the role of topic and seek to understand the underlying dimensions that drive readers’ frame selections.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. “State of the News Media 2014: Pew Research Center’s Journalism Project,” The Pew Research Center, http://www.journalism.org/2014/03/26/state-of-the-news-media- 2014-overview/ (accessed September 1, 2014); Angilee Shah, “Health Journalism Takeaways from the Annual State of the Media Report,” Reporting on Health, March 25, 2011, http://www.reportingonhealth.org/blogs/health-journalism-takeaways-annual-state- media-report (accessed August 23, 2014). 2. Diana C. Mutz, Hearing the Other Side: Deliberative versus Participatory Democracy (Cambridge, UK: Cambridge University Press, 2006); Joyce Y. M. Nip, “Exploring the Second Phase of Public Journalism,” Journalism Studies 7 (April 2006): 212-36. 3. Renita Coleman, Esther Thorson, and Lee Wilkins, “Testing the Effect of Framing and Sourcing in Health News Stories,” Journal of Health Communication 16 (October 2011): 941-54; Avery Holton, Na Yeon Lee, and Renita Coleman, “Commenting on Health: A Framing Analysis of User Comments in Response to Health Articles Online,” Journal of Health Communication 19 (7, 2014): 825-37. 4. Ashley A. Anderson, Dominique Brossard, Dietram A. Scheufele, Michael A. Xenos, and Peter Ladwig, “The ‘Nasty Effect’: Online Incivility and Risk Perceptions of Emerging Technologies,” Journal of Computer-Mediated Communication 19 (3, 2014): 373-87. 5. Lori Dorfman, Lawrence Wallack, and Katie Woodruff, “More than a Message: Framing Public Health Advocacy to Change Corporate Practices,” Health Education & Behavior 32 (June 2005): 320-36; Joan W. Higgins, P. J. Naylor, Tanya Berry, Brian O’Connor, and David McLean, “The Health Buck Stops Where? Thematic Framing of Health Discourse to Understand the Context for CVD Prevention,” Journal of Health Communication 11 (April/May, 2006): 343-58; John McManus and Lori Dorfman, “Functional Truth or Sexist Distortion? Assessing a Feminist Critique of Intimate Violence Reporting,” Journalism 6 (February 2005): 43-65. 6. Coleman, Thorson, and Wilkins, “Testing the Effect of Framing and Sourcing.” 7. Coleman, Thorson, and Wilkins, “Testing the Effect of Framing and Sourcing.” 8. David Domingo, “Managing Audience Participation,” in Participatory Journalism: Guarding Open Gates at Online Newspapers, ed. Jane B. Singer, David Domingo, Ari

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Suran et al. 737

Heinonen, Alfred Hermida, Steve Paulussen, Thorsten Quandt, Zvi Reich, and Marina Vujnovic (Oxford, UK: Wiley-Blackwell, 2011), 9-37. 9. Ari Heinonen, “The Journalist’s Relationship with Users,” in Participatory Journalism: Guarding Open Gates at Online Newspapers, ed. Jane B. Singer, David Domingo, Ari Heinonen, Alfred Hermida, Steve Paulussen, Thorsten Quandt, Zvi Reich, and Marina Vujnovic (Oxford, UK: Wiley-Blackwell, 2011), 33-55; Oscar Westlund, “Transforming Tensions: Legacy Media towards Participation and Collaboration,” Information, Communication & Society 15 (6, 2012): 789-95. 10. Andy Williams, Claire Wardle, and Karin Wahl-Jorgensen, “‘Have They Got News for Us?’ Audience Revolution or Business as Usual at the BBC?” Journalism Practice 5 (1, 2011): 85-99. 11. Anderson et al., “The ‘Nasty Effect.’” 12. Coleman, Thorson, and Wilkins, “Testing the Effect of Framing and Sourcing”; Holton, Lee, and Coleman, “Commenting on Health.” 13. Pamela J. Shoemaker, Philip R. Johnson, Hyunjin Seo, and Xiuli Wang, “Readers as Gatekeepers of Online News: Brazil, China, and the United States,” Brazilian Journalism Research 6 (1, 2011): 55-77. 14. Shanto Iyengar, Is Anyone Responsible?: How Television Frames Political Issues (Chicago: University of Chicago Press, 1991); Robert M. Entman, “Framing: Toward Clarification of a Fractured Paradigm,” Journal of Communication 43 (December 1993): 51-58; Werner J. Severin and James W. Tankard Jr., Communication Theories: Origins, Methods, and Uses in the Mass Media (NY: Addison Wesley Longman, 2001). 15. Katherine W. Hawkins and Darren L. Linvill, “Public Health Framing of News Regarding Childhood Obesity in the United States,” Health Communication 25 (November/December 2010): 709-17; James A. Levine, “Poverty and Obesity in the US,” Diabetes 60 (November 2011): 2667-68; Lesa Hatley Major, “Break It to Me Harshly: The Effects of Intersecting News Frames in Lung Cancer and Obesity Coverage,” Journal of Health Communication 14 (March 2009): 174-88; Brian L. Quick and Benjamin R. Bates, “The Use of Gain- or Loss-Frame Messages and Efficacy Appeals to Dissuade Excessive Alcohol Consumption among College Students: A Test of Psychological Reactance Theory,” Journal of Health Communication 15 (September 2010): 603-28. 16. Iyengar, Is Anyone Responsible?; Entman, “Framing: Toward Clarification of a Fractured Paradigm.” 17. Hatley Major, “Break It to Me Harshly.” 18. Hawkins and Linvill, “Public Health Framing.” 19. Coleman, Thorson, and Wilkins, “Testing the Effect of Framing and Sourcing”; Levine, “Poverty and Obesity in the US”; Hatley Major, “Break It to Me Harshly.” 20. Holton, Lee, and Coleman, “Commenting on Health.” 21. Zvi Reich, “User Comments,” in Participatory Journalism: Guarding Open Gates at Online Newspapers, ed. Jane B. Singer, David Domingo, Ari Heinonen, Alfred Hermida, Steve Paulussen, Thorsten Quandt, Zvi Reich, and Marina Vujnovic (Oxford, UK: Wiley- Blackwell, 2011), 96-117. 22. Coleman, Thorson, and Wilkins, “Testing the Effect of Framing and Sourcing.” 23. Reich, “User Comments.” 24. Hyunyi Cho and Franklin J. Boster, “Effects of Gain versus Loss Frame Antidrug Ads on Adolescents,” Journal of Communication 58 (September 2008): 428-46; Quick and Bates, “The Use of Gain- or Loss-Frame Messages and Efficacy Appeals.”

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 738 Journalism & Mass Communication Quarterly 91(4)

25. Cho and Boster, “Effects of Gain versus Loss Frame”; Hatley Major, “Break It to Me Harshly”; Daniel J. O’Keefe and Jakob D. Jensen, “The Relative Persuasiveness of Gain- Framed Loss-Framed Messages for Encouraging Disease Prevention Behaviors: A Meta- analytic Review,” Journal of Health Communication 12 (October/November 2007): 623-44. 26. Holton, Lee, and Coleman, “Commenting on Health.” 27. Avery Holton, Brooke Weberling, Christopher E. Clarke, and Michael J. Smith, “The Blame Frame: Media Attribution of Culpability about the MMR–Autism Vaccination Scare,” Health Communication 27 (October 2012): 690-701; Brooke W. McKeever, “News Framing of Autism: Understanding Media Advocacy and the Combating Autism Act,” Science Communication 35 (April 2013): 213-40. 28. Reich, “User Comments.” 29. Lesa Hatley Major and Renita Coleman, “Source Credibility and Evidence Format: Examining the Effectiveness of HIV/AIDS Messages for Young African Americans,” Journal of Health Communication 17 (May/June 2012): 515-31. 30. Tanya R. Berry, Joan Wharf-Higgins, and P. J. Naylor, “SARS Wars: An Examination of the Quantity and Construction of Health Information in the News Media,” Health Communication 21 (1, 2007): 35-44. 31. Rachel Smith, “Media Depictions of Health Topics: Challenge and Stigma Formats,” Journal of Health Communication 12 (April/May 2007): 233-49. 32. Jennifer Manganello and Nancy Blake, “A Study of Quantitative Content Analysis of Health Messages in US Media from 1985 to 2005,” Health Communication 25 (July/ August 2010): 387-96. 33. Berry, Wharf-Higgins, and Naylor, “SARS Wars”; Manganello and Blake, “A Study of Quantitative Content Analysis”; Smith, “Media Depictions of Health Topics.” 34. Hatley Major, “Break It to Me Harshly.” 35. Jane Taggart, Anna Williams, Sarah Dennis, Anthony Newall, Tim Shortus, Nicholas Zwar, Elizabeth Denny-Wilson, and Mark F. Harris, “A Systematic Review of Interventions in Primary Care to Improve Health Literacy for Chronic Disease Behavioral Risk Factors,” BMC Family Practice 13 (June 2012), doi:10.1186/1471-2296-13-49. 36. Marie-Claire Shanahan, “Changing the Meaning of Peer-to-Peer? Exploring Online Comment Spaces as Sites of Negotiated Expertise,” Journal of Science Communication 9 (February 2010): 1-13. 37. “HealthNewsReview.org—Independent Expert Reviews of News Stories,” HealthNewsReview.org, www.healthnewsreview.org (accessed August 23, 2014). 38. Daniel Riffe, Charles F. Aust, and Stephen R. Lacy, “The Effectiveness of Random, Consecutive Day, and Constructed Week Sampling in Newspaper Content Analysis,” Journalism & Mass Communication Quarterly 70 (March 1993): 133-39. 39. Douglas A. Luke, Charlene A. Caburnay, and Elisia L. Cohen, “How Much Is Enough? New Recommendations for Using Constructed Week Sampling in Newspaper Content Analysis of Health Stories,” Communication Methods and Measures 5 (1, 2011): 76-91. 40. This study sought to analyze the relationship between the content in online articles and resulting comments. Recognizing that comments can be in response to other comments, and that such interaction can confound the association between story content and comment, only original comments were included in the sample. 41. Smith, “Media Depictions of Health Topics.” 42. Manganello and Blake, “A Study of Quantitative Content Analysis.” 43. Berry, Wharf-Higgins, and Naylor, “SARS Wars.”

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 Suran et al. 739

44. Iyengar, Is Anyone Responsible?; Entman, “Framing: Toward Clarification of a Fractured Paradigm.” 45. Adrian Edwards, Glyn Elwyn, Judith Covey, Elaine Matthews, and Rolsin Pill, “Presenting Risk Information: A Review of the Effects of Framing and Other Manipulations on Patient Outcomes,” Journal of Health Communication 6 (January 2001): 61-82. 46. Andrew F. Hayes and Klaus Krippendorff, “Answering the Call for a Standard Reliability Measure for Coding Data,” Communication Methods and Measures 1 (1, 2007): 77-89. 47. Daniel Riffe, Stephen Lacy, and Frederick G. Fico, Analyzing Media Messages: Using Quantitative Content Analysis in Research (Mahwah, NJ: Lawrence Erlbaum, 2005). 48. For a thorough explanation of point-biserial correlation analysis, see James D. Brown, Understanding Research in Second Language Learning: A Teacher’s Guide to Statistics and Research Design (Cambridge, UK: Cambridge University Press, 1988). 49. Coleman, Thorson, and Wilkins, “Testing the Effect of Framing and Sourcing”; Holton, Lee, and Coleman, “Commenting on Health.” 50. Cho and Boster, “Effects of Gain versus Loss Frame.” 51. Belle B. Cooper, “Why Positive Encouragement Works Better than Criticism, According to Science,” Buffer, January 13, 2014, http://blog.bufferapp.com/why-positive-encouragement-works-better-than-criticism-according-to-science (accessed January 18, 2014). 52. The authors would like to thank an anonymous reviewer for these insights. 53. Norman Sartorius and Hugh Schulze, Reducing the Stigma of Mental Illness: A Report from a Global Association (Cambridge, UK: Cambridge University Press, 2005). 54. Craig Hemmens, Milo Miller, Velmer S. Burton Jr., and Susan Milner, “The Consequences of Official Labels: An Examination of the Rights Lost by the Mentally Ill and Mentally Incompetent Ten Years Later,” Community Mental Health Journal 38 (April 2002): 129- 40; Russell E. Shain and Julie Phillips, “The Stigma of Mental Illness: Labeling and Stereotyping in the News,” in Risky Business: Communicating Issues of Science, Risk, and Public Policy, ed. Lee Wilkins and Phillip Patterson (Westport, CT: Greenwood Press, 1991), 61-74; Otto E. Wahl, Amy Wood, and Renee Richards, “Newspaper Coverage of Mental Illness: Is It Changing?” Psychiatric Rehabilitation Skills 6 (1, 2002): 9-31. 55. Dorfman, Wallack, and Woodruff, “More than a Message”; Higgins et al., “The Health Buck Stops Where?” 56. Dorfman, Wallack, and Woodruff, “More than a Message.” The authors would like to thank an anonymous reviewer for this insight. 57. Jane Harden, “Language, Discourse and the Chronotope: Applying Literary Theory to the Narratives in Health Care,” Journal of Advanced Nursing 31 (March 2000): 506- 12; O’Keefe and Jensen, “The Relative Persuasiveness of Gain-Framed Loss-Framed Messages.”

Downloaded from jmq.sagepub.com at NATL TAIWAN UNIV of Sci and Tech on December 20, 2014 JMQXXX10.1177/1077699014550094Journalism & Mass Communication QuarterlyMaurantonio 550094research-article2014

Myth and History

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 740–755 Remembering Rodney King: © 2014 AEJMC Reprints and permissions: Myth, Racial Reconciliation, sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550094 and Civil Rights History jmcq.sagepub.com

Nicole Maurantonio1

Abstract On June 17, 2012, journalists reported the death of Rodney King, the black motorist whose 1991 beating by several white Los Angeles police officers was captured on video by citizen journalist George Holliday. This essay argues that journalistic mythologizing of Rodney King as a victim of circumstance and journalism as simultaneous hero echoed existing narratives of civil rights history that largely strip black people of agency. In so doing, journalists proffered a larger cultural narrative of racial reconciliation and progress, while recoding King’s life in accordance with other pre-existing, racialized scripts.

Keywords myth, race, Rodney King, civil rights history

On June 17, 2012, Rodney King was found at the bottom of his swimming pool in Rialto, California, and later pronounced dead. King was 47. Without the results of an autopsy or visible signs of foul play, journalists reported King’s death as an accidental drowning. News of Rodney King’s death circulated widely, however, for reasons unlinked to the circumstances surrounding his untimely demise. Rodney King’s place within history was secured on March 3, 1991, when King, a black motorist, was bru- tally beaten by several white Los Angeles police officers following a car chase. The videotaped beating, recorded by white bystander George Holliday, was played and replayed by news media, making the footage “one of the most watched pieces of amateur video in history.”1 The 1991 beating and the 1992 Los Angeles riots that erupted in the wake of the acquittal of the four officers accused in the case made Rodney King

1University of Richmond, Richmond, VA, USA Corresponding Author: Nicole Maurantonio, Department of Rhetoric and Communication Studies, University of Richmond, 28 Westhampton Way, 402-I Weinstein Hall, Richmond, VA 23173, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Downloaded from http://www.elearnica.ir Maurantonio 741

an icon, according to the Reverend Al Sharpton, of the “anti-police brutality and anti- racial profiling movement of our time” and consequently an international celebrity.2 Celebrity deaths are now a familiar part of the journalistic landscape, adhering to pre-existing scripts and, in some cases, rising to the status of so-called “media events.”3 When eulogizing celebrities, particularly celebrities with troubled pasts, journalists tend to adopt, as Kitch has written, a common narrative arc:

an unhappy or very difficult childhood, exceptional talent or beauty, the lucky break or “discovery,” genius or beauty misunderstood, surrender to temptation followed by public disfavor and midlife crisis, recovery and comeback, and the ironic cruelty of death when the person was being appreciated anew.4

Reports of the “sad” and “difficult” life of Rodney King largely conformed to this pattern.5 Journalists highlighted King’s struggles with drug and alcohol abuse (made public during his later appearances on reality television shows such as Celebrity Rehab with Dr. Drew Pinsky in 2008 and Sober House in 2009), his brutal beating immortalized on video in 1991, the simplistic brilliance of his infamous question asked on the third day of rioting in 1992, “Can we all get along?” and the publication of his 2012 memoir The Riot Within: From Rebellion to Redemption, in which Rodney King pro- fessed his forgiveness of the officers who beat him. King’s, according to journalists, was a life cut tragically short. While news coverage of Rodney King’s death reflected a similar sort of narrative repair that, Kitch argues, was evident in the eulogizing of so-called “dark” celebrities—public figures whose lives were fraught with personal demons6—coverage of the death of Rodney King revealed more than simply a familiar narrative emerging anew. Building upon Lule’s definition of myth as a subset of news stories that “offer sacred, societal narratives with shared values and beliefs, with lessons and themes, and with exemplary models that instruct and inform,”7 this essay examines journalistic constructions of two particular mythical narratives: the myth of Rodney King as hapless, helpless victim and the myth of journalism, specifically citizen journalism, as hero. In crafting and circulating these two interconnected myths, journalists located King’s life within existing narratives of civil rights history that largely strip black people of agency, placing the power to enact change within white hands.8 Together, these mythical narratives offer a site to consider the racial dimensions of journalistic constructions of the past. If “news coverage of a death is partly the retelling of the story of a life,”9 as Kitch has argued, in retelling the story of Rodney King’s life, journalists laid the groundwork for a larger cultural narrative of national redemption and racial reconciliation—a narrative in which journalism celebrated its place within historical record. In what follows, this essay first provides additional background for this case and explores journalism’s role in formulating myth as a mode of narrating the past. Next, it discusses journalists’ construction of King as mythical victim, with particular focus upon the “accidental” circumstances surrounding his victimhood. The word “accident” is used in quotes to suggest that references to King’s life experiences as

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 742 Journalism & Mass Communication Quarterly 91(4)

“accidents” reflect a discursive politics that implicitly codes race through more subtly circulating scripts surrounding the black male life. The essay then discusses the myth of (citizen) journalism’s heroism and concludes by reflecting upon the implications of these myths for contemporary U.S. society.

Background and Conceptual Framework The video of Rodney King’s 1991 beating was, Fiske argued, “a discursive instance in which racial and economic antagonism was both expressed through bodily violence and pushed into hypervisibility.”10 As Fiske wrote, “The video[s] were not just video witnesses, technological extensions of the eyewitness, but became video accusers, video defendants, and video verdicts.”11 Holliday’s video of Rodney King’s beating, alongside its various re-technologized versions, raised the question of what it meant to bear witness, an issue, according to Peters, which rests at the core of communication studies.12 For journalism, this question was particularly acute, as the stance of eyewitness is one of the positions from which journalism legitimates its professional authority.13 The question of what it meant to bear witness to the Rodney King beating was further complicated by the Reginald Denny beating, which occurred hours after the 1992 verdict acquitting the officers accused of beating King. The Denny beating was also videotaped. Taken together, as Swenson argued, news media presented the beatings suffered by King and Denny in a manner that “restore[d] racial inequality as social order.”14 According to Swenson, “That the video of black youths assaulting [white] Reginald Denny was chosen to represent the riots reflects the inferential racism intrin- sic to media representations.”15 While journalism fulfilled its larger social function in preserving order, it had also reified hegemonic whiteness, calling the viewer to identify with the white victim, Denny.16 The image of the Rodney King beating, as Spratt argued, “provided the central element of a mediated morality play”—a story pitting good against evil. Yet, like the meaning of the images emerging from the South during the Civil Rights Movement of the 1950s and 1960s featuring clashes between predominantly black civil rights protesters and white police, the meaning of the videotaped beating of Rodney King proved unstable, subject to interpretive revision.17 While Spratt contends that photographs of police dogs attacking protesters taken in 1963 Birmingham, Alabama, “have come to represent and explain . . . the valiant struggle for civil rights by a people who were suppressed, segregated, and abused by the dominate [sic] White power structure,”18 the path to securing that narrative was not inevitable. Rather, journalists played a pivotal role in shaping American mythology.19 Journalists’ stories surrounding Rodney King’s beating functioned similarly, relying on cultural narratives of victims and heroes, injustice and forgiveness. The death of Rodney King little more than twenty years after his infamous beating offered an opportunity to reflect not only upon the life of a man whose identity was publicly defined by the March 3, 1991, beating, but also, once again, upon journalism’s role within society. Rodney King’s death was a story that necessitated reflection upon a traumatic past of racial violence—a past that implicated journalism. When

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Maurantonio 743

confronted with tragedies or crises, incidents that tend to challenge their authority, journalists attempt to reinforce the lines of their community.20 This can take a number of forms. As Berkowitz and Gutsche argued, “Boundary work that strengthens journalism’s cultural authority has included attempts by journalists to reinforce the ties of their own community, to reset the original standards of the field, and to rebuild public confidence for journalism’s social role.”21 Such strategies often invoke the past, locat- ing events within historical contexts or providing historical analogies.22 In considering news coverage in the wake of Rodney King’s death, this essay focuses upon a specific form of storytelling that “reassures by telling tales that explain baffling or frightening phenomena and provide acceptable answers”:23 myth. A particular subset of news stories, myths, however, are more than simply familiar cultural stories helping the public access and process trauma. For journalists, as Berkowitz wrote, “applying myth becomes a believable way of adding narrative coherence while also helping to get the job done on time and in an acceptable form.”24 In addition to their practical function in facilitating journalists’ work, myths provide primers on citizenship and insight into the structures upon which they rely, because myths, according to Bird and Dardenne, can exist “only if they are communicated.”25 It is in the telling and retelling of stories that myths acquire their significance, allowing the values and beliefs central to society to persist.26 Of the seven so-called “master myths” that Lule identifies, two in particular are the focus in this essay: the victim and the hero. Myths of victims, according to Lule, “attempt to reconcile people to the vagaries of human existence—to cruel fate, to bizarre happenstance, to death itself.”27 Such myths “[o]ffer reconciliation,” elevat[ing] and transform[ing] death into sacrifice.”28 The story of the hero, as Lule argued, “help[s] define—greatness.”29 The death of Rodney King, however, not only engaged journalism’s role in shaping mythical narrative and its turn to the past. Given Rodney King’s particular place within civil rights history and the history of police brutality in the United States, journalists were also confronted with a story that necessitated a discussion of race. The “question of race,” which Butler argued followed the 1992 trial of the officers accused in the beating of Rodney King, lays bare a field that is “not neutral.” Such a field, Butler argued, “is itself a racial formation, an episteme, hegemonic and force- ful.”30 The hegemonic force of racially charged myths was evidenced in Lule’s examination of news coverage of the death of Huey Newton, co-founder of the Black Panther Party. Coverage of the controversial Newton downplayed the Black Panther’s historical significance by situating him as a threat to the social order. Newton’s life was narrativized as a cliché, according to Lule, and Newton was cast as a mythological scapegoat, one whose dissent was delegitimized.31 In June 2012, however, none of the stories covering the death of Rodney King suggested that King was a dissenter. Nor was King framed as posing a threat to the LAPD (Los Angeles Police Department)—the script that dominated the courtroom in 1992. Rather, journalists cast Rodney King as an unwitting victim of police violence. King was a victim of “bad luck” who found himself in the wrong place at the wrong time—a victim of history. In adhering to this script, journalists advanced other racialized and

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 744 Journalism & Mass Communication Quarterly 91(4) racist scripts—specifically scripts of the incompetent and incapacitated black man, which Jackson argued, are cataloged among historic projections of black masculinity that work to limit the discursive field of vision.32 According to Schudson, myths “do not tell a culture’s simple truths so much as they explore its central dilemmas.”33 The stories surrounding Rodney King’s life and journalism’s role in it placed into sharp relief the unresolved and largely unspoken racial dimensions of these mythical narratives.

Method To explore the mythical narratives circulating around Rodney King’s life, focusing upon their racial dimensions in recent U.S. history, a textual analysis was conducted of news coverage during the two weeks following the announcement of Rodney King’s death. This textual analysis was guided by three research questions: First, how was race articulated (if at all) in the myths surrounding King’s life? Second, how did journalists’ use of myth work to locate journalism within historical record? Third, what does the circulation of these myths suggest about journalism’s role in shaping public discourse surrounding race and racism in the United States? Evidence was drawn from a Lexis-Nexis search of U.S. newspaper and wire reports based upon the search terms Rodney King. Sources were collected between June 17, 2012, the day King’s death was announced, and July 1, 2012, by which time King’s funeral had taken place. The search for “Rodney King” yielded 194 sources over the two-week period, which were narrowed to 73 of the most relevant for closer, in-depth analysis. After collecting these materials, the sources were searched for prominent themes and patterns in coverage. Analyzing these textual sources over the course of the two weeks following King’s death drew out the discursive themes dominating news coverage. Particular attention was paid to the ways in which journalists attempted to cast King’s life and legacy through the framing of his victimization and journalism’s heroic role in presenting King’s beating to the public. Journalists’ modes of articulating these mythic narratives are discussed, by first examining the construction of King as an “accidental” and “hapless” victim and then journalists’ construction of citizen journalist Holliday and, by extension, journalism as heroic.

Rodney King: The “Accidental” Victim Rodney King’s was a life book-ended by his so-called “accidents.” The features of King’s life, as reported by journalists, however, were tethered to the first “accident” to befall Rodney King—the March 3, 1991, beating, which came after King led Los Angeles police on a high-speed chase through the city. While stories announcing King’s death identified him by the 1991 beating, they tended to conclude by recount- ing King’s troubled life in the wake of the riots. Journalists noted King’s multiple arrests, mostly for misdemeanors connected to substance abuse, and his status as a reality TV star, appearing in the aforementioned programs Celebrity Rehab and Sober

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Maurantonio 745

House. In 2012, the then-celebrity King published his memoir. Just prior to his death, Rodney King conducted a number of media interviews reflecting on the twentieth anniversary of the Los Angeles riots and publicizing his newly released memoir, preaching forgiveness. Recalling the 1991 beating, journalists emphasized the physical violence King endured. In these stories, King was a brutalized object, acted upon by Los Angeles police. He was not only a battered victim, but also someone who found himself in the wrong place at the wrong time. This emphasis, upon the beating as mere “accident,” situated King as a victim of circumstance in March 1991. In the words of Alfred Doblin, King was “a product of bad genes, bad judgment, bad policing and bad luck.”34 Bill Coffin’s reflections followed a similar line of thought,

Some people are born to greatness. Other people have the conditions for greatness thrust upon them. And some people get caught in the crossfire between the two and find themselves in over their head. Such was the sad case of Rodney King, a man who might have lived a relatively mundane life were it not for the extraordinary events that made him a household name, an icon for modern American race relations and his own worst enemy.35

The Los Angeles Times reflected upon King’s death and the lives of the key players in King’s beating, “Gates, Bradley, and Christopher made this city’s history; it merely happened to King.”36 Not only was King “unlucky” in the most literal sense, but he was also a tragic, pitiable figure to whom history “happened.” Rodney King was an icon whose iconic- ity and symbolism were unintentional, confronted with infamy he neither wanted nor was prepared to handle—catapulted unwillingly into history. He was, as one journalist wrote, “a private and wary individual who was reluctantly dragged into the world’s biggest spotlight.”37 Rodney King was not unlike other icons of the civil rights movement, such as Emmett Till, the fourteen-year-old African American boy killed by two white men in Money, Mississippi, in 1955, whose story found itself repeated in journalistic meditations on Rodney King’s life. History, journalists asserted, happened to both Rodney King and Emmett Till, much like it “happened” to Rosa Parks. According to Leonard Pitts Jr., “history” “chose [emphasis added] quiet, dignified Rosa Parks as the emblem of the fight against segregation.”38 History, Pitts claimed, “chose [emphasis added] handsome, prankish Emmett Till as the face of racial violence.” 39 What went largely unacknowledged by journalists discussing these historic African American men and women was that not only was Rosa Parks a very deliberate icon, an activist who carefully strategized her place within the civil rights movement, but also, Rodney King’s beating was not simply the product of “bad luck.” By framing King’s experiences as a consequence of luck, or lack thereof, journalists largely elided the issue of race that lay at the root of King’s experiences. Although journalists referred to King as a “symbol of civil rights,” implying significance to racial discourse in the United States, by dubbing King’s beating a product of misfortune rather than racism, the beating was brutality that could have been avoided if the circumstances had been

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 746 Journalism & Mass Communication Quarterly 91(4) different. Those circumstances, however, did not include institutional racism. In mythologizing King as the consummate victim, who led a “sad” and “difficult” life,40 there was something almost inevitable about Rodney King’s experiences. He was, according to the myth, a sort of pathetic figure who, by coincidence, kept falling into bad situations. By extension, even Rodney King’s struggles with substance abuse could largely be written off as another piece in an otherwise troubled life, beset by tragedy.

Rodney King: The “Hapless” Victim Although coding Rodney King as an accidental victim inverted the script of King as violent threat that had gained traction during the 1992 trial, journalists advanced con- currently circulating scripts coding black masculinity. Journalists’ repeated reference to the “difficult” life Rodney King led in the wake of the 1991 beating—his inability to maintain a job or his strained relationships with his family—reinforced scripts of incompetence and ignorance.41 These scripts worked in tandem to fortify the myth of Rodney King as “hapless” victim, serving to reinforce it in two ways. First, journalists reflected upon King’s infamous question, “Can we all get along?” as a moment of simplistic brilliance, spoken from the mouth of a man who did not fully understand the weight of his words. Second, journalists reflected, somewhat incredulously, upon King’s ability to forgive those who beat him. Journalists asserted King not only as unthreatening, but also as an unwitting victim who provided fodder for elaborated narratives of national progress with respect to racism in the United States.

Can We All Get Along? When it was asked to television cameras on the third day of deadly rioting in Los Angeles, Rodney King’s infamous question was posed by a man described as “visibly shaken by what was going on.”42 Begging the city for peace, King, according to the San Francisco Chronicle, showed signs of speech impairment due to the beating. He looked “lost and scared,” said the Intelligencer Journal.43 “Plaintive words spilled out of him: “Can we all get along?” For the Intelligencer Journal, the significance of those words was not lost. “Another man named King, a generation before, had articulated the same thought, with the preacher’s rhetorical flourishes and full orchestration. Rodney King put it in six syllables.”44 The comparison between Rodney King and Martin Luther King Jr. suggests their collective goal of advancing the Civil Rights Movement and significance in catalyzing change. However, Rodney King’s question bore none of the markers of polish and rehearsal that defined Martin Luther King’s oratory. While, as Georgetown Professor Michael Eric Dyson was quoted, “You ain’t got to be highly educated and deeply connected to a bourgeois black infrastructure to make a statement that articulates and summarizes the hopes, aspirations, dreams and determination of a people,”45 the sense of surprise intoned by journalists suggests that Dyson’s point was largely not part of the discursive framing of Rodney King’s infamous question. Rather than grant legitimacy to Rodney King’s experiences and the

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Maurantonio 747

words those experiences spawned, journalists once again played upon the unintentional brilliance of the question. Jeff Yang of CNN dubbed the question “inadvertently profound.”46 Much like Rodney King’s status as accidental victim, his powerful words, too, were unplanned—accidental— affording King little agency. Such discussions of King pointed to his general “haplessness,” as Leonard Pitts Jr. of the Miami Herald wrote. Pitts commented in a widely republished piece,

But there is a reason Shakespeare put wisdom into the mouths of fools.

The fool could get away with saying what others could not.

No, King was not a fool. But he was a hapless guy, taken less than seriously—in part because he asked that question others would not. Yet that question, the one some of us tried to giggle into irrelevance, is the defining question of the American experiment.47

Pitts’s commentary ultimately concluded by asking, “But if the man who believes we must all get along is a fool, then you really have to wonder: what word is left for the man who does not?” In posing this question, Pitts afforded King greater credit than most, calling readers to carefully consider the so-called “foolishness” of the question. Yet, the construction of King as “hapless” bears the markers of journalistic constructions of King’s “accidental” and seemingly pitiable existence. Pitts argued that when asking for calm amid chaos, Rodney King was “unforgivably earnest” in asking his question. There was, he wrote, “something guileless, naked, even innocent” in King’s question. It was “a signature moment, when Rodney Glen King was not hapless.”48 King was positioned as a man whose childlike naiveté accidentally, albeit powerfully, spoke to the nation. He wielded little agency even over the words he spoke. A similar construction of King as incompetent and “hapless” was articulated by Dennis McDougal, who wrote, “He [Rodney King] struck me as a sort of Baby Huey character—basically sweet-natured but always blundering into one disaster after another.”49 Within this framing, King was well intentioned, but lacking the where- withal to take control of his life or to understand the gravity of his place in history. As the San Gabriel Valley Tribune argued,

. . . as with almost everything else in his life, the fact that he was engaged to a juror in his civil trial would, if written into a novel or screenplay, seem terribly unrealistic. That was Rodney King, who never sought much, but ended up as both a Zelig-like figure in history and possessed of an almost Buddha-like calm, sought out for comments on race relations whenever anything went wrong, or right.50

Reflecting upon King as “Zelig-like” suggested that no one really knew the “real” Rodney King. He was a construction—mediated and arguably defined by those around him. He was little more than the product of others’ decisions, others’ narratives. John Burris, the Oakland civil rights attorney who represented King in his civil trial, sum- marized, “He became sort of a trophy, who was passed around and placed in situations that he really wasn’t equipped to handle.”51

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 748 Journalism & Mass Communication Quarterly 91(4)

He “Turned His Scars into Stars” King’s willingness to forgive was described as yet another indicator of his childlike naiveté and general “haplessness.” Across coverage of King’s funeral, journalists repeatedly commented on King as “a forgiving man who bore the scars of his infamous beating with dignity.”52 The repetition of this phrase across news outlets nationwide suggests that King was to be lauded for bearing the burden of his beating in such a quiet way, making him a “model” civil rights icon. The burden, of course, was not simply physical. It was historical—reminiscent of the brutality enacted by whites against black people for centuries. Despite this, Rodney King was not angry. He did not turn the injustice of American racism into a site for expressing anger. Rather Rodney King, as the Reverend Al Sharpton reminded, “became a symbol of forgiveness.” “Rodney had risen above his mistakes. He never mocked anyone, not the police, not the justice system, not anyone.”53 Lawrence Spagnola, who co-authored King’s memoir, reiterated Sharpton’s claim by stating that King’s family could be proud of the “‘amazing degree of grace and wisdom’ with which King carried himself after being violently thrust into the media spotlight.”54 Within his 2012 memoir, Rodney King further fueled the discursive frame of forgiveness, describing himself as a man who deemed forgiveness a way of granting simultaneous closure to the past and hope for the future. King claimed,

My country’s been good to me . . . This country is my house, it’s the only home I know, so I have to be able to forgive—for the future, for the younger generation coming behind me so . . . they can understand it and if a situation like that happened again, they could deal with it a lot easier.55

In reiterating the words of King alongside Sharpton’s suggestion that King “turned his scars into stars,” journalists placed repeated emphasis upon the importance of forgiveness, transferring King’s legacy from one of reminding Americans of the necessity of confronting racism and police abuse of authority to the forgiveness of those aforementioned sins. King did what white America could only hope: Rodney King had forgiven the nation for its racism. A moment of national racial reconciliation, it seemed unthinkable given the brutality King endured. The victimized King’s forgiveness may be the one site where he was granted some agency in the narrative surrounding his place within civil rights history. However, his forgiveness was framed as action incred- ible, almost inexplicable, and once again reflective of an almost childlike resilience.

(Citizen) Journalism: The Hero If Rodney King’s forgiveness was an indicator of the nation’s racial reconciliation—its forgiveness for past racism—the opportunity to reflect upon Rodney King’s life, too, offered journalists a moment to reflect upon journalism’s role in U.S. civil rights history. The 1991 Rodney King beating was, as Tom Marquandt of the Capital wrote, “a key moment in citizen journalism.”56 Marquandt sought to remind readers that the name of Rodney King would be little more than a footnote in U.S. history had it not

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Maurantonio 749

been for George Holliday’s amateur video. Without the bystander Holliday’s video, Marquandt reminds, “there never would have been a case against police officers—or so much fury about their acquittal.”57 Alicia Stewart of CNN.com posited that part of the power of the video of the King beating stemmed from the fact that it came from an “unexpected source: a citizen journalist.”58 Jesse Washington of the Associated Press reiterated, “If a man had not stepped outside of his home and videotaped the beating, King would have been lost to history.”59 Journalism was that which made history. King was simply there—again in the wrong place at the wrong time. Although King’s beating was described by journalists as the product of “bad timing,” as discussed earlier in this paper, little was made of Holliday’s so-called “good timing.” Discussion of Holliday’s place within history seemed to suggest that Holliday’s role, the white bystander poised to show the beating to the world, was a given. The death of Rodney King thus was a reminder of not only King’s place within history—civil rights history, urban history, and media history—but also journalism’s place within historical record. As Lynn Elber of the Associated Press reminded,

It had long been the job of TV news cameras to document history and inform members of the public, who could serve as eyewitnesses but not reporters and certainly not camera operators. The most noteworthy exception: Abraham Zapruder, the Dallas businessman who recorded the critical seconds of President John F. Kennedy’s 1963 assassination on 8 MM film.60

Los Angeles Civil Rights Association president Eddie Jones claimed, during an event held to memorialize Rodney King, “The good part is there was somebody there with audio and video and technology. That has made the difference in the way things are being run today in law enforcement.”61 Much like the Zapruder film presents the final moments of President Kennedy’s life, the Holliday video of Rodney King’s beating offers the only documentation of the violence. Both instances, however, were immortalized not by professional journalists, but by members of the public. Both Zapruder and Holliday are critical reminders of citizen journalism’s centrality to history and historical record. By championing Holliday’s video and Holliday as a hero, a sentiment repeated by the Reverend Jesse Jackson in public statements in the wake of King’s death, it was the tape of Rodney King’s beating that made “[t]he problem of excessive force in American policing real,”62 as if to suggest that the mediated videotape produced a verifiable truth that previously was dubious. Reference to Rodney King, some twenty-plus years following the beating, conjures the powerful memory of Holliday’s video. The visual has been seared into American memory, as Emily Langer of the Washington Post wrote.63 In under- scoring the power of the video, agency was transferred to the medium, which allowed the public to bear witness to the action taken against King. Rodney King was once again stripped of his place as an historical agent. George Holliday was not the only hero. Journalism, more broadly defined, had performed heroically, as journalists reminded. Just as the video stands as testament to the power and necessity of (citizen) journalism, so too did journalistic coverage of King’s death elicit broader journalistic

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 750 Journalism & Mass Communication Quarterly 91(4) reflection upon journalists’ authority as eyewitnesses to the 1992 verdict acquitting the officers charged with beating King and the riots that followed. Tammerlin Drummond of the Bay Area News Group reflected upon the days of rioting in the wake of the 1992 verdict and, specifically, the dissension in the newsroom. Drummond, however, wrote,

Yet in the midst of this horrible time, there were acts of extraordinary kindness and courage. An African-American colleague at the Times risked his life to rescue a Vietnamese woman who had been attacked by a mob. The riots brought out the worst in many, but others rose to the occasion.64

Although the Los Angeles riots are remembered as among the deadliest riots in American history, they were a time that also signaled that good still existed, particularly among journalists. Reflection upon Rodney King’s life offered not simply an opportunity to consider citizen journalism’s significance; it offered an opportunity to reconsider journalism’s institutional role more broadly within the citizenry.

Conclusion While journalists’ mythologizing of Rodney King as an unwitting and accidental victim rather than aggressor reversed the narrative that had gained traction during the 1992 trial, their coding of King as “hapless” served to reinforce other dominant scripts racially schematizing the black male life: scripts of incapacitation and incompetence. Although King’s significance within recent U.S. history is indisputable, the profundity of his words, like his suffering, were cast by journalists as inadvertent, unintentional. Yet, the “accidents” that defined King’s life were not without purpose, or so journalists claimed. Rather, King’s life was told as a triumphant narrative for understanding race relations in the United States. In journalists’ telling, King suffered brutality unimagi- nable so that (white) audiences across the United States could see what African Americans had claimed was institutionalized for centuries: racism. As Bill Coffin wrote,

And while the violence of his own beating and the larger violence of the riots that followed shocked King and the nation, one might say that in the end, the nation emerged from it stronger, healthier, and better able to face the future.65

The lesson had been learned, or so journalists posited. Mythologized in this way, Rodney King’s life was transformed into a medium for national healing. Police departments specifically and the nation more broadly were redeemed. The Daily News of Los Angeles declared the commemoration of the twentieth anniversary of the riots and meditations upon the life and death of Rodney King a reminder of

how far the LAPD has come since the era of the King beating. A larger police force seems to have given officers the confidence to enforce the law without overstepping ethical

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Maurantonio 751

bounds. And policy changes mean officers can no longer expect it to be covered up if they do step over the line.66

The story of King’s life was thus a story of redemption—of moving beyond the traumatic past and learning the lessons of history. Rodney King had sacrificed his life for the good of the nation. Like the stories of many victims before him, who, as Lule wrote, were “guilty only of coincidence, bad timing, and the unfortunate fate of being in the wrong place at the wrong time,”67 Rodney King’s story was resolved. Yet few would remember the name of Rodney King had it not been for the video of a citizen journalist who left an indelible imprint on American racial consciousness. The video had provided tangible evidence—proof—that was indisputable. Journalism provided the nation with the ability to bear witness to the brutality suffered by King. Journalism had fulfilled its most basic function. Thus, in remembering Rodney King as the consummate victim, one who was himself a construction, journalists portrayed Rodney King as a man who ceded control over his fate to those who would shape public memory of his life and legacy. In so doing, journalists shaped their own legacy within civil rights history as members of the profession who exposed police brutality for the world to witness—as heroes. Retelling the story of Rodney King’s “accidental” and “hapless” life thus allowed journalism to engage in a form of self-celebration that reiterated the significance of journalism to racial discourse in the United States. King’s story, of course, is not the first of its kind. In narrating the life of Rodney King and invoking the names of historical figures such as Emmett Till and Rosa Parks, journalists summoned a past intimately connected to journalism. The photographs of Till’s mutilated body were published by Jet magazine in 1955 and cited as a catalyst of the modern civil rights movement.68 The images of Rosa Parks seated on the Montgomery bus where she refused to move to the “colored” section are, too, inscribed within popular memory as sparking the movement for civil rights in the United States. Rodney King’s story was situated in relation to these narratives—narratives in which journalism played a central role, as it did in Birmingham, Alabama, where Charles Moore’s photographs captured police dogs attacking black protesters under the eyes of Bull Connor.69 In these cases, the black men and women attacked were portrayed by journalists as wielding little authority. They were framed as passive, thereby relying upon sympathetic witnesses to rescue them from their plight.70 Journalists’ remem- brances of Rodney King conformed to pre-circulating histories of the civil rights movement. However, while the story of the Rodney King beating conformed to existing civil rights histories, journalists’ retelling of Rodney King’s life served as a redemptive coda—reminding of not only King’s forgiveness for the nation’s sins of racism, but also of the purpose of Rodney King’s life, which enabled the nation to see changes that needed to be made and, according to journalists, were. Journalism’s authority was once again reaffirmed and King’s life narrativized as a national parable, albeit once again schematized through a racist lens, following Butler and Jackson.71 The myths of King as victim and journalism as hero bear consequence for understandings of how journalistic mythologizing can reinforce racism and narratives of history that privilege whiteness. When Rodney King died in June 2012, the story of the

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 752 Journalism & Mass Communication Quarterly 91(4) slain Florida teen Trayvon Martin was starting to unfold, once again bringing to the fore questions of racism and the appropriation of police power. While names like Emmett Till once again made their way into popular discourse, so, too, were commen- taries arguing that things had changed since the time of the Rodney King beating. Philadelphia talk radio host Dom Giordano was quoted, “For instance like a Trayvon Martin, I do see things routinely that indicate that we are getting along, that we are moving past racial tensions.”72 Although changes were made in response to the video of Rodney King’s beating, there is danger in moving too far toward narratives that suggest national redemption and closure, as the recent George Zimmerman acquittal reminds. Rodney King’s life and death are testament to the unfinished, unresolved past. The controversy surrounding Trayvon Martin’s death is yet another reminder that the narrative is more complicated and troubled than journalists’ stories grant. Although myths provide journalists with scripts within which to fit stories, the limits to mythical narrative are, as the case of Rodney King and subsequent similar cases remind, profound. The myth of Rodney King’s life had been resolved, preserving the social order. Myth’s potential to change the social order, a more difficult proposition as Lule suggests, however, remains unrealized.73

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. Bill Coffin, “In Passing: Rodney King (1962-2012),” National Underwriter Life & Health/Financial Services, July 1, 2012, http://www.lifehealthpro.com/2012/07/01/ rodney-king-1962-2012. 2. Jesse Washington, “Rodney King’s Plea Measures His Lasting Meaning,” Associated Press, June 17, 2012, http://bigstory.ap.org/article/rodney-kings-plea-measures-his-lasting-meaning. 3. Carolyn Kitch, “‘It Takes a Sinner to Appreciate the Blinding Glare of Grace’: Rebellion and Redemption in the Life Story of the ‘Dark’ Celebrity,” Popular Communication 5 (1, 2007): 37-56. On “media events,” see Daniel Dayan and Elihu Katz, Media Events: The Live Broadcasting of History (Cambridge, MA: Harvard University Press, 1992). 4. Kitch, “It Takes a Sinner to Appreciate the Blinding Glare of Grace,” 38. 5. “Sudden, Sad End to a Difficult Life,” Inland Valley Daily Bulletin (Ontario, CA), June 17, 2012, http://www.dailybulletin.com/20120618/sudden-sad-end-to-a-difficult-life. 6. Kitch, “It Takes a Sinner to Appreciate the Blinding Glare of Grace,” 38. 7. Jack Lule, Daily News, Eternal Stories: The Mythological Role of Journalism (NY: Guilford, 2001), 18. 8. Martin A. Berger, Seeing through Race: A Reinterpretation of Civil Rights Photography, with foreword by D. J. Garrow (Berkeley: University of California Press, 2011).

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Maurantonio 753

9. Carolyn Kitch, “‘A Death in the American Family’: Myth, Memory, and National Values in the Media Mourning of John F. Kennedy Jr.,” Journalism & Mass Communication Quarterly 79 (2, 2002): 299. 10. John Fiske, Media Matters: Race and Gender in U.S. Politics (Minneapolis: University of Minnesota Press, 1996), 126. 11. Fiske, Media Matters, 126. 12. John Durham Peters, “Witnessing,” Media, Culture & Society 23 (6, 2001): 707-23. 13. Barbie Zelizer, Covering the Body: The Kennedy Assassination, the Media, and the Shaping of Collective Memory (Chicago: University of Chicago Press, 1992), 131. 14. Jill Dianne Swenson, “Rodney King, Reginald Denny, and TV News: Cultural (Re-) Construction of Racism,” Journal of Communication Inquiry 19 (1, 1995): 86. 15. Swenson, “Rodney King, Reginald Denny, and TV News,” 80. 16. On journalism’s role in preserving social order, see Herbert J. Gans, Deciding What’s News: A Study of CBS Evening News, NBC Nightly News, Newsweek, and Time (NY: Vintage Books, 1979), 52-62. For more on “encoding the beating of Reginald Denny,” see Swenson, “Rodney King, Reginald Denny, and TV News,” 81-84. 17. Meg Spratt, “When Police Dogs Attacked: Iconic News Photographs and Construction of History, Mythology, and Political Discourse,” American Journalism 25 (2, 2008): 86. 18. Spratt, “When Police Dogs Attacked,” 102. 19. Spratt, “When Police Dogs Attacked.” 20. Barbie Zelizer, “Journalists as Interpretive Communities,” Critical Studies in Mass Communication 10 (3, 1993): 219-37; Dan Berkowitz and Robert E. Gutsche, “Drawing Lines in the Journalistic Sand: Jon Stewart, Edward R. Murrow, and Memory of News Gone By,” Journalism & Mass Communication Quarterly 89 (4, 2012): 643-56; Dan Berkowitz, “Doing Double Duty: Paradigm Repair and the Princess Diana What-a-Story,” Journalism 1 (2, 2000): 125-43. 21. Berkowitz and Gutsche, “Drawing Lines in the Journalistic Sand,” 644. 22. Jill A. Edy, “Journalistic Uses of Collective Memory,” Journal of Communication 49 (2, 1999): 71-85. 23. S. Elizabeth Bird and Robert W. Dardenne, “Myth, Chronicle and Story: Exploring the Narrative Qualities of News,” in Social Meanings of News, ed. Dan Berkowitz (Thousand Oaks, CA: SAGE, 1997), 336. 24. Dan Berkowitz, “The Ironic Hero of Virginia Tech: Healing Trauma through Mythical Narrative and Collective Memory,” Journalism 11 (6, 2010): 648. 25. Bird and Dardenne, “Myth, Chronicle and Story,” 337. 26. Bird and Dardenne, “Myth, Chronicle and Story,” 337. 27. Lule, Daily News, Eternal Stories, 43. 28. Lule, Daily News, Eternal Stories, 22. 29. Lule, Daily News, Eternal Stories, 23. 30. Judith Butler, “Endangered/Endangering: Schematic Racism and White Paranoia,” in Reading Rodney King/Reading Urban Uprising, ed. Robert Gooding-Williams. (New York: Routledge, 1993), 17. 31. Lule, Daily News, Eternal Stories, 62-64. 32. Ronald L. Jackson II, Scripting the Black Masculine Body: Identity, Discourse, and Racial Politics in Popular Media (Albany: State University of New York Press, 2006), 75. 33. Michael Schudson, cited in Kitch, “It Takes a Sinner to Appreciate the Blinding Glare of Grace,” 40.

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 754 Journalism & Mass Communication Quarterly 91(4)

34. Alfred P. Doblin, “Karen Klein Takes a Ride into Celebrity,” The Record (Bergen County, NJ), June 25, 2012, p. A11. 35. Coffin, “In Passing: Rodney King (1962-2012).” 36. The Los Angeles Times cited in “What others say: Rodney King a major part of Los Angeles modern history,” Deseret Morning News, June 20, 2012, http://www.deseretnews. com/article/765584574/Rodney-King.html?pg=all 37. Jeff Yang, “Opinion: Can We All Get Along?” CNN: InAmerica Blog, June 20, 2012, http:// inamerica.blogs.cnn.com/2012/06/20/opinion-can-we-all-get-along/. 38. Leonard Pitts Jr., “Rodney King Was a Hapless Icon,” State Journal-Register (Springfield, IL), June 25, 2012, p. 6. 39. Pitts, “Rodney King was a Hapless Icon.” 40. “Sudden, Sad End to a Difficult Life.” 41. Jackson, Scripting the Black Masculine Body. 42. Coffin, “In Passing: Rodney King (1962-2012).” 43. In Our View, “Words of Wisdom,” Intelligencer Journal/New Era (Lancaster, PA), June 27, 2012, p. A10. 44. In Our View, “Words of Wisdom,” p. A10. 45. Washington, “Rodney King’s Plea Measures His Lasting Meaning.” 46. Yang, “Opinion: Can We All Get Along?” 47. Pitts, “Rodney King Was a Hapless Icon,” p. 6. 48. Pitts, “Rodney King Was a Hapless Icon,” p. 6. 49. Douglas McDougal, “Rodney King: Legacy of a Flawed Martyr,” Tulsa World, June 21, 2012, p. A17. 50. In Our View, “Words of Wisdom,” p. A10. 51. Bruce Newman, “Oakland Attorney John Burris Remembers Tough Life of Rodney King, ‘The Face of Police Brutality,’” San Jose Mercury News, June 18, 2012, http://www.mer- curynews.com/ci_20879447/oakland-atty-john-burris-remembers-tough-life-rodney. 52. “Rodney King Remembered at Funeral as Forgiving Man,” USA Today, June 30, 2012, http:// usatoday30.usatoday.com/news/nation/story/2012-06-30/rodney-king-funeral/55948254/1. 53. “Rodney King Remembered at Funeral as Forgiving Man.” 54. “Rodney King Remembered at Funeral as Forgiving Man.” 55. CNN Wire Staff, “Friends, Family Remember Rodney King at Funeral,” CNN Wire, July 3, 2012, http://www.cnn.com/2012/07/01/us/rodney-king-funeral/. 56. Tom Marquandt, “These Days, the Candid Camera Is Always On,” The Capital (Annapolis, MD), June 24, 2012, p. 12. 57. Marquandt, “These Days, the Candid Camera Is Always On,” p. 12. 58. Alicia W. Stewart, “5 Ways the Rodney King Beating and LA Riots Changed America,” CNN: InAmerica Blog, June 18, 2012, http://inamerica.blogs.cnn.com/2012/06/18/5- ways-the-rodney-king-beating-and-la-riots-changed-america/. 59. Washington, “Rodney King’s Plea Measures His Lasting Meaning.” 60. Lynn Elber, “Rodney King Beating Helped Drive Video Revolution,” Huffington Post, June 18, 2012, http://www.huffingtonpost.com/2012/06/18/rodney-king-video-of- beat_n_1607177.html. 61. Gabriel L. Acosta, “Memorial Held at Rodney King Residence in Rialto,” San Bernardino Sun, June 22, 2012 http://www.sbsun.com/general-news/20120623/memorial- held-at-rodney-king-residence-in-rialto. 62. Stewart, “5 Ways the Rodney King Beating and LA Riots Changed America.”

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 Maurantonio 755

63. Emily Langer, “Rodney King Dies; Victim of L.A. Police Beating Was 47,” Washington Post, June 17, 2012, http://articles.washingtonpost.com/2012-06-17/ national/35462044_1_police-officers-george-holliday-police-brutality. 64. Tammerlin Drummond, “Rodney King Death Brings Back Memories of L.A. Riots,” Inside Bay Area, June 18, 2012 http://www.insidebayarea.com/ci_20885166/tammerlin- drummond-rodney-king-death-brings-back-memories. 65. Coffin, “In Passing: Rodney King (1962-2012).” 66. Editorial, “Just the Facts: Public Should Learn What Happened in Police Shootings,” The Daily News of Los Angeles, June 20, 2012, p. A12. 67. Lule, Daily News, Eternal Stories, 54. 68. Christine Harold and Kevin Michael DeLuca, “Behold the Corpse: Violent Images and the Case of Emmett Till,” Rhetoric & Public Affairs 8 (2, 2005); 263-86; Brian Thornton, “The Murder of Emmett Till: Myth, Memory, and National Magazine Response,” Journalism History 36 (2, 2010): 96-104. 69. Berger, Seeing through Race; Davi Johnson, “Martin Luther King Jr.’s 1963 Birmingham Campaign as Image Event,” Rhetoric & Public Affairs 10 (1, 2007): 1-25. 70. Berger, Seeing through Race, 9-57. 71. Butler, “Endangered/Endangering,” and Jackson, Scripting the Black Masculine Body. 72. Washington, “Rodney King’s Plea Measures His Lasting Meaning.” 73. Lule, Daily News, Eternal Stories, 193.

Downloaded from jmq.sagepub.com at UNIVERSIDAD DE CONCEPCION on December 20, 2014 JMQXXX10.1177/1077699014550089Journalism & Mass Communication QuarterlyHuntsberger 550089research-article2014

Myth and History

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 756–771 Attempting an Affirmative © 2014 AEJMC Reprints and permissions: Approach to American sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550089 Broadcasting: Ideology, jmcq.sagepub.com Politics, and the Public Telecommunications Facilities Program

Michael W. Huntsberger1

Abstract The Public Telecommunications Facilities Program (PTFP) was the largest source of capital funding for U.S. public broadcasters for nearly fifty years. Between 1963 and 2010, the PTFP distributed more than $800 million to support the construction of public broadcasting facilities. Though the PTFP itself was generally noncontroversial, the fortunes of the program were complicated by the partisan politics of public broadcasting and federal fiscal policy. This study provides evidence of the ambiguous and contingent nature of the American approach to public broadcasting, and demonstrates some of the problems associated with affirmative efforts by government to advance public communication.

Keywords history, law and policy, public broadcasting

In many parts of the world, public broadcasting is an enduring national presence in the lives of citizens. Agencies such as the British Broadcasting Corporation in the United Kingdom and Nippon Hoso Kyokai (NHK) in Japan are embedded in the structures of their parent democracies and in the patterns of people’s daily lives. In the United States, however, public broadcasting has not achieved a similar status. America’s more limited efforts were first associated with investments in education.1 The most visible

1Linfield College, McMinnville, OR, USA Corresponding Author: Michael W. Huntsberger, Linfield College, 900 SE Baker St., A587, McMinnville, OR 97128, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Downloaded from http://www.elearnica.ir Huntsberger 757

outcome of this movement was the passage of the Public Broadcasting Act of 1967 and the subsequent establishment of the Corporation for Public Broadcasting (CPB), which remains the primary support agency for public radio and television in the United States.2 However, the CPB was not the only agency established to encourage American public broadcasting, or even the first one. From 1963 to 2010, the Public Telecommunications Facilities Program (PTFP) played a significant role in bringing radio and television services to millions of Americans. Primarily as an agency of the Department of Commerce (DOC), the PTFP distributed more than $800 million in matching funds to noncommercial educational broadcast licensees in all fifty states. The PTFP was the largest single source of capital funding for public broadcasting, helping local stations to build and acquire transmission facilities, studios, remote broadcasting services, satellite operations, and other equipment, as well as the infrastructure to house, interconnect, and operate them. The leverage provided by PTFP awards impelled universities, colleges, school districts, indigenous nations, and nonprofit groups to raise hundreds of millions of dollars in their local communities to support projects that extended the reach of public broadcasting, and enhanced the relationships that bind content providers to their audiences. But unlike the CPB, financial assistance from the PTFP was available to every noncommercial, non-sectarian broadcaster in the United States. The PTFP was part of President John F. Kennedy’s aspiration to achieve social progress through public sector initiatives. The program persisted through decades of political turbulence and tug-of-war. What began as an effort to assist a small set of undercapitalized TV stations became part of a systematic endeavor to extend the reach of broadcasting in America. In the end, the mass communication technologies facili- tated by the PTFP were eclipsed by the intersection of newer technologies, transforma- tive circumstances, and intractable ideological conflicts over the proper role of government in the broadcasting sector.

Literature Review and Research Questions The First Amendment prohibits government interference in free expression, but not government assistance. Emerson catalogs a series of circumstances where government has acted to promote public communication to compensate for “major distor- tions in the system” and “the failure of the market place of ideas to operate according to the original plan.”3 Comparing the electromagnetic spectrum with the infrastructure of public highways, Emerson asserts that the state has an obligation to make communication channels “available to potential participants in the system, both communicators and listeners.”4 Emerson includes public broadcasting among “certain types of governmental participation, such as education and public libraries,” that are intended to “promote expression on the part of others.”5 Specifically, Emerson finds that, “as economic and technical developments make access to the media of communication more difficult for individuals or groups without financial resources,” the “voluntary furnishing of facilities by the government”6 is necessary and appropriate.

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 758 Journalism & Mass Communication Quarterly 91(4)

In the 1920s, advocates for noncommercial and educational interests articulated this affirmative view, as debate proceeded over how the emerging radio broadcasting system might best serve the interests of the American public. The affirmative view was generally opposed by advocates of limited government and commercial interests, who favored approaches to broadcasting that relied on private enterprise. This “instrumental view,” according to Avery and Stavitsky, equates “the interest of the public with that of the industry being regulated.”7 McChesney documents how advocates for commercial broadcasting systematically marginalized the interests of educators, religious institutions, labor unions, and other noncommercial agencies during broadcasting’s formative years.8 Between 1928 and 1935, the United States adopted an instrumental approach to broadcast regulation that heavily favored commercial, nationally distributed operators. McChesney points to the Federal Radio Commission’s General Order 40, and the subsequent reallocation of more than 90% of the radio spectrum in a manner favorable to commercial interests, as impediments to the development of European- style public service broadcasting in the United States.9 Led by commissioner and public service advocate Frieda Hennock, the Federal Communications Commission (FCC) acknowledged the affirmative view in 1948, by reserving a portion of the broadcast spectrum exclusively for educational broadcasting stations. While this action prompted a reconsideration of the role of government in public expression through educational broadcasting, the debate proceeded at a glacial pace. It was another fifteen years before Congress authorized a program to provide federal support for noncommercial broadcasting.10 At the time, a system of educational television stations was envisioned as the response to the “vast wasteland” of commercial television content criticized by FCC Chair Newton Minow.11 This view held that commercial television catered to the mass audience with lowest common denominator entertainment programs, and that noncommercial media would provide content of higher quality that could uplift the tastes and sensibilities of that audience. Scholars of cultural studies have questioned this position. Oullette asserts that the American system of public broadcasting reflects an elitist view of commercial broadcasting, conflating “commercial hegemony with unquestioned assumptions about the social, cultural, and moral inadequacies of ‘mass appeal.’”12 Such charges mirror one of the fundamental challenges to public broadcasting in the political sphere. President Richard Nixon raised charges of elitism against the public affairs programming on public television when he vetoed appropriations to the Corporation for Public Broadcasting in 1972. In Engleman’s view, Nixon sought to “dismantle the newly established public broadcasting system,” based on his belief “that public television had a left/liberal agenda hostile to his administration.”13 In the 1980s, the assault expanded to include the PTFP, when the program was swept up in the efforts of the administration of Ronald Reagan to defund public broadcasting. For advocates of government investments in public telecommunications facilities, controversies over content were secondary. By providing matching support for local facilities projects, not content, the PTFP embraced the fundamental value of localism enshrined in broadcast policy since the 1920s. This differentiation has largely gone unnoticed by scholars. For example, Tressel, Buckelew, Suchy, and Brown lumped the

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Huntsberger 759

PTFP together with the larger public broadcasting system, describing both as a “historical accident” in a 1977 report for the U.S. Office of Education.14 While the authors correctly observed that American public broadcasting reflected “the difficulty of reducing broad intangible goals to practical policy,”15 they overlooked the goals and objectives that were explicitly articulated in the documents on the PTFP. The current study examines these documents in detail. Lacking the mandate of consensus, public broadcasting’s advancement (or obstruc- tion) has often depended on the involvement of key figures in the government. Krasnow and Longley describe Presidents Kennedy and Nixon as “actively interested in broadcast matters.”16 In the legislative branch, members of Congress often play equally pivotal roles in public communications policy, they say, “because of their seniority or their influential standing in a committee.”17 Krasnow and Longley specifically acknowledge the influence of Senator Warren Magnuson, a Democrat and chair- person of the Senate Commerce Committee, whose word was “practically law at the FCC.”18 As one of the earliest proponents of educational television, and a sponsor of the Educational Broadcasting Facilities Act of 1962, Magnuson was instrumental in the establishment of the PTFP and the American public broadcasting system.19 The PTFP received varying degrees of attention in some of the early histories of the U.S. public broadcasting. Burke traces the origins of the PTFP to a meeting of the board of the National Association of Educational Broadcasters (NAEB) in Atlanta.20 Gibson documents the establishment of Educational Broadcasting Facilities Program (EBFP), the precursor to the PTFP, and observes that the program moved through the Congress with “very little opposition.”21 Blakely recognizes the EBFP as the initiative that “to a large extent shaped public broadcasting” as the precursor to CPB.22 Mitchell offers a detailed investigation of the political history of the PTFP, covering the period from 1963 through 1985. Her historical study demonstrates how changes in the PTFP reflected adjustments in political priorities and attitudes toward federal funding. Supported by the evidence presented in primary sources from the legislative and executive branches of government, Mitchell asserts that the PTFP persisted across the first two decades because of its “noncontroversial nature, its geographical reach, and its ability to attract money from state, local, and private entities.”23 At the same time, the study describes continuous, quiet opposition to the program for its ongoing costs, its redundancy, and its minor role in the larger public broadcasting system. After Mitchell’s study, the PTFP received little scholarly attention. Avery attributes the lack of sustained scholarly focus on public broadcasting to the demise of the NAEB, because the organization served as “the principal scholarly intersection for broadcast educators and professional broadcasters who were following careers in instructional, educational, or public broadcasting.”24 In Avery’s view, the working arrangements between educational broadcasters and their parent institutions grew more distant, as the educational broadcasting outlets of the 1950s evolved into the public broadcasting system of the 1980s, and each focused on core missions and services. The present investigation provides closure to the investigations of the PTFP undertaken by Mitchell and others, and closes the gap in the literature of public broadcasting

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 760 Journalism & Mass Communication Quarterly 91(4) with respect to this particular policy initiative. The present research examines the con- trasting views of government involvement in the media throughout the entire history of the PTFP. The study analyzes the forces at work in an expansive and sustained public initiative supported by the political system. The examination relies on historical methods to address three research questions:

RQ1: What happens when elected federal officials take an affirmative role in the establishment and propagation of a robust public telecommunications system? RQ2: What were the arguments for and against the PTFP? RQ3: What circumstances and conditions contributed to the demise of the PTFP?

The research relies on archival reports, memoranda, and other materials from agencies of the executive and legislative branches dating from 1965 to 2011. Appropriation figures were compared with reports in other documentary sources. Additional sources include papers stored in the National Public Broadcasting Archive at the University of Maryland, as well as interviews with officers of the PTFP and key figures within the public broadcasting industry. While these interviews provided little direct evidence to the study, they provided context, and confirmed the centrality of the PTFP to the achievement of a national system of public radio and television stations.

Support for Educational Broadcasting Facilities The PTFP began as an idea of NAEB legal counsel Leonard Marks.25 In 1956, Marks proposed that the NAEB advocate federal legislation to provide funds for the development of local educational television (ETV) stations in the United States. At the time, only twenty-four ETV stations were operating on the 242 channels set aside four years earlier by the FCC for noncommercial TV broadcasting.26 With the nation facing growing school enrollments and shortages of facilities, Marks believed educators could successfully make the case for addressing the nation’s educational objectives for children and adults through the new technology of television. Marks was the communications attorney for Texas Senator Lyndon Johnson, a Democrat, whose background as an educator and a broadcast owner sensitized him to the goals and objectives of educational broadcasters and the NAEB.27 Senator Johnson played a pivotal role in moving the ETV initiative through Congress. The NAEB ultimately achieved passage of the Educational Television Facilities Act in 1962. At the outset, the federal program to support educational broadcasting was of limited scope and scale. The legislation charged the Department of Health, Education, and Welfare (DHEW) to offer grants to support “the construction of educational television broadcasting facilities.”28 The act authorized $32 million over five years for projects applied for by June 30, 1968. Though Congress debated limits on eligibility, the list of qualifying agencies in the act allowed for a wide range of applications from public primary and secondary schools, public institutions of higher education, private colleges and universities supported in part by tax revenues (such as national defense research

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Huntsberger 761

grants), or nonprofit agencies eligible for noncommercial educational television broadcast licenses and “organized primarily to engage in or encourage educational television.”29 The act limited awards to up to 50% of the “reasonable and necessary cost of such a project,”30 plus 25% of the costs of facilities owned by the applicant. The act limited total awards to any single state to $1 million. The act instructed the Secretary of the Department of Health, Education, and Welfare to develop rules and policies for the program, including the priority order for awards. Finally, the act responded to concerns about federal influence by prohibiting agencies or employees of the federal government from exercising any control over ETV facilities, or “over the curriculum, program of instruction, or personnel of any educational institution, school system, or educational broadcasting station or system.”31 Though President Kennedy signed the act on May 1, 1962, no spending occurred until a modest $1.5 million appropriation became available in fiscal year 1963 to the Educational Television Facilities Program (ETFP). In the first year, five awards total- ing $858,952 were given to recipients in Illinois, South Carolina, Virginia, and Utah. The following year, the program received more applications, and awards were given to sixteen new ETV stations and the expansion of another ten existing facilities.32 The ETFP received $13 million for fiscal year 1965, followed by smaller appropriations in the succeeding years until the originally authorized amount was disbursed in its entirety. With a landslide victory in 1964, a clear mandate, and majorities in both houses of Congress, President Lyndon Johnson was a powerful advocate for big government policies and programs. As part of his vision for the Great Society, he championed the creation of a national system of educational television and radio stations.33 Signed into law on November 7, 1967, the Public Broadcasting Act extended the authority of the ETFP through fiscal year 1971, and included steadily increasing authorizations for $10.5 million, $12.5 million, and $15 million in each of the following three fiscal years. The act replaced the $1 million per-state limit with a flat 8.5% per-state cap, extended grants to U.S. territories, collapsed the provisions for the maximum federal match to 75% for any award, and expanded the program to include grants to cover the costs of facility planning initiatives.34 The act, however, contained no provision for long-term financing for public broadcasting, thus avoiding possible constitutional challenges associated with government involvement in matters of free expression.35 With the expanded mandate to serve both television and radio, the ETFP was renamed the Educational Broadcasting Facilities Program (EBFP) within the DHEW. The program was nested within the department’s Office of Education as one of several federal programs supporting adult education. But alongside continuing education, public library services, and initiatives in the arts and humanities, the EBFP was an outlier. Opinions varied in the agency and in the administration about the proper role of the facilities awards and the relationship of the projects with those being undertaken by the CPB. Even as the Nixon administration accused public broadcasters of political bias, the facilities program remained relatively noncontroversial because it was more closely identified with locally autonomous stations. Though Nixon vetoed CPB

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 762 Journalism & Mass Communication Quarterly 91(4) financing in 1972, touching off a battle with Congress that continued for months,36 EBFP appropriations remained untouched. The program was extended through 1975 in Public Law 93-84, and again for another two years by President Ford in the Educational Broadcasting Facilities and Telecommunications Demonstration Act of 1976.37 President Carter’s budget for fiscal year 1979 recommended transferring the EBFP to the CPB to resolve the apparent redundancy, and included a corresponding increase in CPB funding.38 Though the plan was supported by DHEW and CPB, congressional conservatives challenged Carter’s proposal. Unlike the relatively noncontroversial infrastructure initiatives, the CPB remained closely associated with politically charged content in public television. In the view of longtime public radio and television manager Dennis Haarsager, EBFP’s specific focus on local infrastructure appealed to key members of Congress in both parties, who saw an opportunity to extend the program to cable, satellite, and other telecommunications technologies for public and educational services. One such legislator was South Carolina Senator Ernest Hollings. As governor, Hollings had worked with public broadcasters in his state to acquire EBFP funds to expand the South Carolina Educational Television network.39 Under the circumstances, it became more politically feasible to keep the program independent of the CPB by transferring it to the National Telecommunications and Information Administration (NTIA), a new agency in the Department of Commerce. The proposal became reality in the Public Telecommunications Act of 1978. The act also reautho- rized the program at a level of $40 million through fiscal year 1981.40

Educational Broadcasting Becomes Public Telecommunications With the move to the NTIA, the EBFP was renamed the Public Telecommunications Facilities Program. The scope of the initiative widened to include nonbroadcast projects in addition to those for planning, broadcasting, and interconnection. Under the rules and policies adopted in 1979, the PTFP gave highest priority to projects that provided “telecommunications facilities for first service to a geographic area,” followed in priority order by projects to activate or extend “significantly different additional services” to geographic areas, projects to improve existing station facilities, and projects to add new capabilities to existing stations.41 Congress appropriated, and President Carter approved $24 million for fiscal year 1980, the highest level of financing since the program’s inception. The new priorities touched off a flurry of comments from noncommercial broadcasting interests, including National Public Radio, the National Federation of Community Broadcasters, and the Moody Bible Institute.42 Across the nation, hopeful broadcasters and planning groups jockeyed for priority status as they identified first- service opportunities in unserved and underserved localities. The political environment of the PTFP changed dramatically with the election of President Ronald Reagan in 1980. The philosophies of the New Deal and the Great

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Huntsberger 763

Society were replaced in the executive branch by a staunchly conservative ideology, and deep commitments to smaller government and lower taxes. Reagan’s initial budget proposed a cut of almost thirty million dollars to the CPB as part of a broader effort to “encourage direct beneficiaries and the private sector to make larger contributions to cultural activities.”43 The budget included no funds for the PTFP after fiscal year 1982.44 Reagan offered no plan to transfer the activities of the PTFP to CPB. Instead, to serve the broader goal of reduced government spending, the program would simply be eliminated. While the budget document offered no justification for zeroing out the program, the prevailing posture of the administration toward public broadcasting implied that the PTFP was considered to be an unnecessary burden on the public trea- sury.45 The 1983 fiscal year proposal was the first of twelve consecutive annual executive budgets that sought to eliminate funding for the PTFP. In March 1982, with the country in the grips of the most sustained economic recession in fifty years, the administration proposed another round of budget reductions. Reagan asked the Congress to rescind the PTFP’s existing authorization and terminate the agency outright. Immediately, the program had to stop processing applications. But once again, the grassroots appeal of the program proved hard to overcome. Local interests lobbied their representatives and senators to stand up for projects in their districts and states. Congress responded by disapproving the administration request. The action allowed PTFP to resume its activities.46 With the immediate crisis averted, a diverse group of legislators, including conservative Republican Senator Barry Goldwater of Arizona and liberal Democratic Representative Timothy Wirth of Colorado, stepped forward with bipartisan plans to ensure the future of the PTFP and reauthorize funding for the coming years. Though Wirth’s attempt to reauthorize the program in a new version of the Public Telecommunications Act failed, some of the provisions of Wirth’s proposal survived in the Budget Reconciliation Act of 1981, including the reauthorization of the PTFP through fiscal year 1984.47 While this made the survival of the program more likely, the subsequent appropriation for fiscal year 1983 was reduced to $15 million. The $12 million appropriated the following year cut the program to half of its size under the Carter administration. Though the authorization lapsed in fiscal year 1985, members of the divided Congress declared administration proposals to zero out the PTFP “dead on arrival,” and managed to restore the program to previous funding levels through a series of continuing resolutions.48 When the broader issue of the federal role in public broadcasting resurfaced in the 100th Congress, the grassroots appeal of the PTFP mani- fested itself again in the Public Telecommunications Act of 1988, which included authorizations for both the PTFP and the CPB through fiscal year 1991.49 As the executive branch passed to the administration of George H. W. Bush, the sharp ideological confrontations that threatened the PTFP in the early 1980s gave way to a less confron- tational period. While the administration continued to advocate the elimination of the PTFP, Congress continued annual appropriations at levels comparable with those established in the decade before. With the passage of the Public Telecommunications Act of 1992, the reauthorization of the program through fiscal year 1994, and the subsequent election of Democrat Bill Clinton, it seemed that the future of the program

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 764 Journalism & Mass Communication Quarterly 91(4) was more secure.50 The 1992 act, however, marked the last time the PTFP was granted the security of a multi-year authorization.

The Long Denouement During the Clinton administration, the PTFP reappeared as a line item in every executive budget. The final appropriation for 1994 restored the program to $24 million. But the following fall, federal politics underwent another dramatic shift, as conservatives took control of both houses of Congress for the first time in four decades, re-awaken- ing the partisan divisions of the early 1980s. Guided by the policy goals of the Reagan years, energized congressional conservatives took up the causes of smaller government, lower taxes, and the privatization of cultural programs, including public broadcasting.51 The political conflict over the federal budget led to a shutdown of non-essential federal agencies for six days in November 1995. Unable to agree on a budget, the Congress resolved the stalemate by a series of continuing resolutions that allowed government agencies to re-open for twelve months. In the aftermath of the battle, the PTFP appropriation was reduced to $13.5 million.52 Following another, longer shutdown the following year, the appropriation was reduced again by continuing resolution to $13 million.53 The PTFP was restored to pre-shutdown levels after Clinton’s 1996 re-election, but continuing divisions in the government made it impos- sible to achieve the compromises required to reauthorize the program.54 Even without an authorization, other actions of the 104th Congress supported the continuation of the PTFP. Under the Telecommunications Act of 1996, television broadcasters were given ten years to convert their transmission systems from analog to digital technologies.55 Subsequently, as part of the Clinton Administration campaign to develop the National Information Infrastructure (NII), the PTFP received appropriation increases between 2000 and 2003 to assist stations with digital transmission conversion.56 The additional funds helped hundreds of public television and radio stations in most U.S. states to upgrade to digital transmission. The promise of digital television service for all Americans, and the momentum of the digital conversion initiative, helped to sustain the PTFP in the Congress, and through the change of administrations in 2001. In fiscal year 2003, the program was again eliminated in the executive budget proposal of Republican President George W. Bush. As before, a combination of grassroots lobbying by public broadcasters and citizen groups and actions by key legislators managed to sustain the PTFP as it was swept up in annual fights over CPB funding. In 2005, with longtime public broadcasting advocate Republican Ted Stevens of Alaska serving as president of the Senate, the PTFP was sustained even after the Republican- controlled House Appropriations Committee voted for its elimination. Public television manager Steve Bass attributed the victory to the efforts of individual constituents, and the leadership of members of Congress.57 Subsequent battles took place with regu- larity in the following years, even as the makeup of the Congress shifted. Stevens of Alaska, Hollings of South Carolina, and other longtime members of the Congress who had worked with public broadcasters to bring PTFP projects to their constituencies were voted out or retired.

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Huntsberger 765

In 2008, the United States was thrust into the largest financial crisis since the Great Depression. PTFP’s project of digital conversion was largely complete, and the long- simmering issues of cost, redundancy, and federal support for social programs returned to the forefront of the debate over public broadcasting. The forces of fiscal conserva- tism unleashed in the aftermath of the crisis proved to be more than any constituency, individual, or group of legislators could overcome. Even the election of Democrat Barack Obama could not stop the momentum toward fiscal austerity. In May 2010, the director of the chief executive’s Office of Management and Budget cited the PTFP as an example of federal spending that was “not merit-based” and “plainly wasteful and duplicative.”58 In February 2011, the Obama administration informed Congress of its intent “to terminate new funds for the Public Telecommunications Facilities Grant Program.”59 Asserting that the CPB was the more appropriate agency to assist capital projects, the proposal allowed the President to appease fiscal conservatives with a budget cut of $18 million. Members of both political parties, in both houses of Congress, accepted the proposal virtually without debate. Retired former PTFP director Dennis Connors commented that the program “was an easy target” in the arena of federal politics.60 In a few trade journals, a handful of public station managers conveyed anxiety about the erosion of federal support, and some engineers expressed concern for the replacement of aging infrastructure,61 but no general outcry arose in the broader press. After nearly half a century, the primary program that helped to build America’s system of public radio and television stations passed quietly into history.

Looking Back on the PTFP The trajectory of the appropriations to the PTFP offers some insight into the history of the program (Figure 1). When the effects of inflation are accounted for, the data show that the largest investment of government resources was made at the outset, when policy makers embraced Johnson’s lofty goals. After the reauthorization of 1967, conservative policy makers attempted to substantially reduce or eliminate the PTFP in fiscal years 1975, 1983, and 1996/1997. Each of these efforts took place at a time when policy makers were engaged in ideological debates about the affirmative role of the federal government in the advancement of a national system of public broadcasting. Though the PTFP was politically popular and noncontroversial, after 1967 the program was regularly hobbled by its association with the CPB, and more fundamentally by the marginal and controversial status of public broadcasting in the United States. The mission of the PTFP was limited to the broadly accepted tasks of supporting the construction of local facilities and the acquisition of equipment, making the program popular with legislators. But the funding was often tied to the political fortunes of a larger and more politically volatile agency, the CPB. The significant increases in PTFP appropriations coincided with key developments in the construction of the public broadcasting system. At each of these points, politically powerful individuals advanced the cause of the PTFP through the political system. With Johnson’s leadership at the outset, the program reinvigorated the nation’s

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 766 Journalism & Mass Communication Quarterly 91(4)

Figure 1. PTFP appropriations/disbursements, 1963-2010; actual versus adjusted for inflation (per U.S. dollar CPI).62 Note. For the purpose of comparison over the forty-eight-year history of the PTFP, appropriations are adjusted with reference to the value of the U.S. dollar in 1963. PTFP = Public Telecommunications Facilities Program; CPI = Consumer Price Index. existing educational stations, and established new ones. In the 1980s, the PTFP was bolstered by the bipartisan actions of Wirth and Goldwater, extending broadcast services into unserved and underserved areas, and linking stations to distribution networks through satellite and other interconnection facilities. In the 1990s, within Clinton’s national telecommunications infrastructure, the PTFP’s digital transition initiatives touched practically every federal district, and provided legislators with an easy justification to overlook the fiscal concerns of the moment to accommodate immediate, one-time projects that benefitted their constituents directly. Blakely and Mitchell give credit for the popularity and success of the PTFP to the contributions of private, local, and state interests. By requiring local participation in projects, the PTFP engendered grassroots investment in and support for the agencies that received PTFP awards. The financial exigencies of 2008 greatly reduced funding and support from these constituencies. State and local government revenues plum- meted. Contributions from private foundations dried up as they lost investment income. Private citizens worried about the economy and reduced donations to chari- table causes. As long as some combination of non-federal resources could be cobbled together to support PTFP projects, federal policy makers had reasons to contribute to the success of the initiatives. After 2008, the grassroots momentum that sustained the PTFP for decades dissipated against the prevailing forces of the recession.

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Huntsberger 767

In the view of NTIA Associate Administrator Bernadette McGuire-Rivera, digital transition was the PTFP’s greatest achievement. But once the transition was complete, there was “no further need for” the PTFP in the view of the Obama administration and the Congress.63 The transition became the program’s endgame. Six decades after the establishment of the PTFP, the CPB counts more than 900 public radio and 350 public television stations in its sphere.64 More stations operate without CPB assistance. These numbers suggest that for policy makers, the mission of the PTFP was substantially fulfilled at the time the program was terminated. Moving forward, the future of ter- restrial broadcasting is intertwined with the regulatory, economic, and technical conditions associated with broadband delivery.65

Conclusion The Public Telecommunications Facilities Program played a significant role in the establishment of public broadcasting in the United States. Though several attempts were undertaken, federal policy makers were never able to agree on a set of principles and policies to establish a more permanent approach to funding the public telecommunications facilities system. Consequently, the PTFP remained a program that served the short-term interests of policy makers and their immediate constituencies. Mired in the conflicts and constraints embedded in the American approach to public broadcasting, the PTFP operated under a narrow directive to encourage investments in the hard- ware and infrastructure associated with homegrown capital projects. As the EBFP, the initiative had connotations of individual empowerment and local control that dovetailed neatly with commitments to localism contained in the Communications Act of 1934. The PTFP skirted around the more controversial issue of direct federal involvement in broadcasting. With the passage of the Public Broadcasting Act of 1967, and the emergence of the CPB, the politics of the PTFP became entangled with those of the larger, more controversial agency. The shift from “educational” to “public” broadcasting represented a significant expansion of the federal government’s role as an affirmative force for public communication. Thereafter, the fortunes of the PTFP were connected to the more erratic politics of the CPB. The ideological battle lines have been apparent from the beginning. When the Public Broadcasting Act of 1967 was passed, minority Republicans in the House led by Samuel L. Devine of Ohio argued that the “economic system is on the verge of collapse and many desirable things must be put aside indefinitely. Could anyone seriously argue that this program to enhance public broadcasting is indispensable?”66 In the early years, the conflict was muted by the noncontroversial rhetoric of public education, and mitigated by the economic buffer of the postwar boom. Over time, calls for fiscal restraint merged with other apprehensions. Even under these challenging circumstances, the PTFP accomplished a series of important objectives, including the build-out of transmission and production facilities, the interconnection of broadcast stations to content distribution systems, and the conversion from analog to digital transmission technologies. Time will tell if a more permanent approach to the American public media system will emerge in the future.

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 768 Journalism & Mass Communication Quarterly 91(4)

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. John Witherspoon, Roselle Kovitz, Robert K. Avery, and Alan G. Stavitsky, A History of Public Broadcasting (Washington, DC: Current, 2000), 13. 2. Witherspoon et al., History of Public Broadcasting, 20-21. 3. Thomas I. Emerson, “The Affirmative Side of the First Amendment” (paper 2768, Faculty Scholarship Series, 1981), 795, The correct URL is http://digitalcommons.law.yale.edu/ fss_papers/2768/ (accessed June 6, 2013). 4. Emerson, “The Affirmative Side of the First Amendment,” 824. 5. Emerson, “The Affirmative Side of the First Amendment,” 800. 6. Emerson, “The Affirmative Side of the First Amendment,” 815. 7. Robert K. Avery and Alan G. Stavitsky, “The FCC and the Public Interest: A Selective Critique of U.S. Telecommunications Policy-Making,” in Public Broadcasting and the Public Interest, ed. Michael P. McCauley, Eric E. Peterson, B. Lee Artz, and DeeDee Halleck (Armonk, NY: M.E. Sharpe, 2003), 53. 8. Robert W. McChesney, Telecommunications, Mass Media, and Democracy: The Battle for the Control of U.S. Broadcasting, 1928-1935 (NY: Oxford University Press, 1994), 1. 9. McChesney, Telecommunications, Mass Media, and Democracy, 25-29. 10. Witherspoon et al., History of Public Broadcasting, 9-11. 11. Laurie Oullette, Viewers Like You? (NY: Columbia University Press, 2002), 17. 12. Oullette, Viewers Like You?, 221. 13. Ralph Engelman, Public Radio and Television in America: A Political History (Thousand Oaks, CA: SAGE, 1996), 168. 14. George W. Tressel, Donald P. Buckelew, John T. Suchy, and Patricia L. Brown, The Future of Educational Telecommunication: A Planning Study (Lexington, MA: D.C. Heath and Company, 1975), 1. 15. Tressel et al., The Future of Educational Telecommunication, 1. 16. Erwin G. Krasnow and Lawrence D. Longley, The Politics of Broadcast Regulation (NY: St. Martin’s Press, 1978), 50. 17. Krasnow and Longley, The Politics of Broadcast Regulation, 70. 18. Krasnow and Longley, The Politics of Broadcast Regulation, 69. 19. W. Wayne Alford, “The Educational Television Facilities Act of 1962,” A-V Communication Review 15 (1, 1967): 84. 20. John E. Burke, An Historical-Analytical Study of the Legislative and Political Origins of the Public Broadcasting Act of 1967 (NY: Arno Press, 1979), 36. 21. George H. Gibson, Public Broadcasting: The Role of the Federal Government, 1912-1976 (NY: Praeger, 1977), 156. 22. Robert J. Blakely, To Serve the Public Interest: Educational Broadcasting in the United States (Syracuse, NY: Syracuse University Press, 1979), 198.

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Huntsberger 769

23. M. Helena Mitchell, “Role of the Federal Government in the Funding of Public Telecommunications Facilities” (PhD diss., Syracuse University, 1987), ii. 24. Robert K. Avery, “Contemporary Public Telecommunications Research: Navigating the Sparsely Settled Terrain,” Journal of Broadcasting & Electronic Media 40 (1, 1996): 132. 25. Burke, An Historical-Analytical Study, 36. 26. Burke, An Historical-Analytical Study, 29. 27. Robert Caro, The Years of Lyndon Johnson: Means of Ascent (NY: Alfred A. Knopf, 1990), xxxiii. 28. Educational Television Facilities Act of 1962, Pub. L. No. 477, 87th Cong., 2nd Sess. (May 1, 1962), Public Broadcasting PolicyBase, Current Newspaper, http://www.current. org/1962/05/educational-television-facilities-act-of-1962-2/ (accessed March 10, 2012). 29. Educational Television Facilities Act of 1962. 30. Educational Television Facilities Act of 1962. 31. Educational Television Facilities Act of 1962. 32. Mitchell, “Role of the Federal Government,” 84. 33. Witherspoon et al., History of Public Broadcasting, 17. 34. Public Broadcasting Act of 1967, Pub. L. No. 129, 90th Cong., 1st Sess. (November 7, 1967), Public Broadcasting PolicyBase, Current Newspaper, http://www.current. org/1967/11/public-broadcasting-act-of-1967-2/ (accessed March 10, 2012). 35. Engelman, Public Radio and Television in America, 159. 36. “Nixon Administration Public Broadcasting Papers—Summary of 1972,” Public Broadcasting PolicyBase, Current Newspaper, http://www.current.org/1979/02/the-nixon- administration-public-broadcasting-papers-summary-of-1972/ (accessed June 21, 2013). 37. Mitchell, “Role of the Federal Government,” 148. 38. Executive Office of the President, Office of Management and Budget, “Budget of the United States Government Fiscal Year 1981” (Federal Reserve Archive, Federal Reserve Bank of St. Louis, January 28, 1980), 217, http://fraser.stlouisfed.org/docs/publications/ usbudget/bus_1981.pdf (accessed January 10, 2013). 39. Les Brown, “Carnegie Report on Public Radio and TV Reflects Panel’s Political Astuteness: News Analysis Intersystem Feuding Bill Being Revised,” New York Times, February 1, 1979, p.C15, http://search.proquest.com.ezproxy.linfield.edu:2048/docview/120998134? accountid=12106 (accessed June 21, 2013). 40. Public Telecommunications Act of 1978, Pub. L. No. 567, 95th Cong., 2nd Sess. (November 2, 1978), Federal Communications Commission, http://transition.fcc.gov/Bureaus/OSEC/ library/legislative_histories/1132.pdf (accessed January 20, 2013). 41. “Public Telecommunications Facilities Program Report and Order Docket No. 78-1,” 44 Federal Register 30,898 (May 29, 1979) (to be codified at 15 C.F.R. pt. 2301). 42. “Public Telecommunications Facilities Program Report and Order Docket No. 78-1.” 43. Executive Office of the President, Office of Management and Budget, “Budget of the United States Government Fiscal Year 1983” (Federal Reserve Archive, Federal Reserve Bank of St. Louis, February 8, 1982), 5-117, http://fraser.stlouisfed.org/docs/publications/ usbudget/bus_1983.pdf (accessed January 10, 2013). 44. Executive Office of the President, Office of Management and Budget, “Budget of the United States Government Fiscal Year 1983,” 8-44. 45. “83 Federal Programs: A Profile of the Reagan Targets,” New York Times, February 29, 1981, http://www.nytimes.com/1981/02/20/us/83-federal-programs-a-profile-of-reagan- targets.html?pagewanted=1 (accessed June 21, 2013).

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 770 Journalism & Mass Communication Quarterly 91(4)

46. “Public Broadcasting Wins Struggles for ’84 Funds,” New York Times, July 20, 1982, http:// www.nytimes.com/1982/07/20/arts/public-broadcasting-wins-struggle-for-84-funds.html (accessed June 21, 2013). 47. Omnibus Budget Reconciliation Act of 1981, Pub. L. No. 35, 97th Cong., 1st Sess. (July 20, 1981), Federal Communications Commission, http://transition.fcc.gov/Bureaus/OSEC/ library/legislative_histories/1172.pdf (accessed March 29, 2012). 48. “Fowler Circulates Memo: FCC, NTIA Budgets Would Increase under Reagan Proposal to Congress,” Communications Daily 6 (25, 1986): 1, http://www.lexisnexis.com.ezproxy. linfield.edu:2048/hottopics/lnacademic/database (accessed March 13, 2012). 49. Public Telecommunications Act of 1988, Pub. L. No. 626, 100th Cong., 2nd Sess. (November 7, 1988), Federal Communications Commission, http://transition.fcc.gov/ Bureaus/OSEC/library/legislative_histories/1354.pdf (accessed March 28, 2012). 50. Public Telecommunications Act of 1992, Pub. L. No. 356, 102nd Cong., 2nd Sess. (August 26, 1992), Federal Communications Commission, http://transition.fcc.gov/Bureaus/OSEC/ library/legislative_histories/1430.pdf (accessed March 28, 2012). 51. Karen DeWitt, “Gingrich Foresees a World without Public Broadcasting,” New York Times, December 17, 1994, p. 9, http://search.proquest.com.ezproxy.linfield.edu:2048/docview/ 109309571?accountid=12106 (accessed June 21, 2013). 52. National Telecommunications and Information Administration, “Annual Report for Fiscal Year 1996,” National Telecommunications and Information Administration, http://www. ntia.doc.gov/legacy/ntiahome/annreports.html (accessed March 27, 2012). 53. National Telecommunications and Information Administration, “Annual Report for Calendar Year 1997,” National Telecommunications and Information Administration, http://www.ntia.doc.gov/legacy/ntiahome/annreports.html (accessed March 27, 2012). 54. National Telecommunications and Information Administration, “Annual Report for Calendar Year 1998,” National Telecommunications and Information Administration, http://www.ntia.doc.gov/legacy/ntiahome/annreports.html (accessed March 27, 2012). 55. Telecommunications Act of 1996, Pub. L. No. 104, 104th Cong., 2nd Sess. (February 8, 1996), Library of the Clerk, U.S. House, http://www.gpo.gov/fdsys/pkg/PLAW- 104publ104/pdf/PLAW-104publ104.pdf (accessed January 21, 2013). 56. Executive Office of the President, Office of Management and Budget, “Budget of the United States Government Fiscal Year 2001” (Federal Reserve Archive, Federal Reserve Bank of St. Louis, February 7, 2000), 140, http://fraser.stlouisfed.org/docs/publications/ usbudget/BUDGET-2001-BUD.pdf (accessed January 10, 2013). 57. Jeremy Enger, “House Votes 2 to 1 to Restore CPB Aid,” Current Newspaper, June 27, 2005, http://www.current.org/wp-content/themes/current/archive-site/cpb/cpb0512fund- ing.shtml (accessed June 21, 2013). 58. “OMB Cites $25 Million to Pubcasting as Example of Unnecessary Spending,” Quick Takes, Current.org, May 25, 2010, http://www.current.org/2010/05/omb-cites-25-million- to-pubcasting-as-example-of-unnecessary-spending/ (accessed June 21, 2013). 59. Executive Office of the President, Office of Management and Budget, “Terminations, Reductions, and Savings, Budget of the U.S. Government, Fiscal Year 2012” (The White House), 42, http://www.whitehouse.gov/sites/default/files/omb/budget/fy2012/assets/trs. pdf (accessed March 25, 2012). 60. Steve Behrens, “CPB Survives, but Not the Facilities Program,” Current Newspaper, April 18, 2011, http://www.current.org/federal/fed1108ptfp.html (accessed March 28, 2012). 61. Randy Stine, “PTFP Shutdown Leaves Pubcasters Scrambling,” Radio World, July 5, 2011, http://radioworld.com/article/ptfp-shutdown-leaves-pubcasters-scrambling/23846 (accessed March 28, 2012).

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 Huntsberger 771

62. Data compiled from Executive Office of the President, Bureau of the Budget, Budget of the United States Government Fiscal Years 1965-1970; Executive Office of the President, Office of Management and Budget, Budget of the United States Government Fiscal Years 1971-2012; Office of Management and Budget, Special Analyses, Budget of the United States Government, Fiscal Years 1971-1980; Office of Management and Budget Analytical Perspectives, 1996-2012, “Federal Reserve Bank of St. Louis,” http://fraser.stlouisfed. org/docs/publications (accessed January 10, 2012); “National Telecommunications and Information Administration Annual Reports, 1994-2000, 2002, 2005, 2006,” National Telecommunications and Information Administration, http://www.ntia.doc.gov/legacy/ ntiahome/annreports.html (accessed March 27, 2012). Inflation adjustments employed the Consumer Price Index Inflation Calculator available from the U.S. Department of Labor Bureau of Labor Statistics. 63. Executive Office of the President, Office of Management and Budget, “Terminations, Reductions, and Savings, Budget of the U.S. Government, Fiscal Year 2012,” 42. 64. “How Many Public Media Stations Are There?” n.d., About Public Media, Corporation for Public Broadcasting, http://www.cpb.org/aboutpb/faq/stations.html (accessed January 25, 2013). 65. Roland Beutler, “The Future Role of Broadcasting in a World of Changing Electronic Communication,” EBU Technical Review (winter 2013): 1, http://tech.ebu.ch/docs/techre- view/trev_2013-Q1_Broadcasting_Beutler.pdf (accessed June 21, 2013). 66. U.S. House, Amendment to the Communications Act of 1934—Report Together with Minority Views and Additional Views (House Report 1281), 90th Cong., 2nd Sess. (April 4, 1968), 7.

Downloaded from jmq.sagepub.com at UNIV CALIFORNIA SAN DIEGO on December 20, 2014 JMQXXX10.1177/1077699014550092Journalism & Mass Communication QuarterlyLiang et al. 550092research-article2014

Communicating Science and Crises

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 772–791 Building Buzz: (Scientists) © 2014 AEJMC Reprints and permissions: Communicating Science in sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550092 New Media Environments jmcq.sagepub.com

Xuan Liang1, Leona Yi-Fan Su1, Sara K. Yeo2, Dietram A. Scheufele1, Dominique Brossard1, Michael Xenos1, Paul Nealey3, and Elizabeth A. Corley4

Abstract Public communication about science faces novel challenges, including the increasing complexity of research areas and the erosion of traditional journalistic infrastructures. Although scientists have traditionally been reluctant to engage in public communication at the expense of focusing on academic productivity, our survey of highly cited U.S. nano-scientists, paired with data on their social media use, shows that public communication, such as interactions with reporters and being mentioned on Twitter, can contribute to a scholar’s scientific impact. Most importantly, being mentioned on Twitter amplifies the effect of interactions with journalists and other non-scientists on the scholar’s scientific impact.

Keywords media and society, communication effects, science communication, social media

For many researchers, communicating with the public about research results rarely entails more than a press release through their institution’s public relations division, and possibly a follow-up interview with a journalist. Only a minority of scientists have been actively engaged in communicating science through popular media outlets. Among them are prominent and highly visible researchers, such as Carl Sagan, Richard

1University of Wisconsin-Madison, Madison, WI, USA 2University of Utah, Salt Lake City, UT, USA 3University of Chicago, Chicago, IL, USA 4Arizona State University, Phoenix, AZ, USA Corresponding Author: Dietram A. Scheufele, Department of Life Sciences Communication, University of Wisconsin-Madison, 309 Hiram Smith Hall, 1545 Observatory Drive Madison, WI 53706, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Downloaded from http://www.elearnica.ir Liang et al. 773

Smalley, and Neil deGrasse Tyson. In spite of these visible exceptions, there continues to be a normative assumption among scientists that public communication is not valuable or is even detrimental to their academic careers.1 Most believe scientists are expected to be modest and dedicated to their research, rather than trumpeting their work in popular media.2 The rewards for communicating science through traditional media are thus believed to compromise a scientist’s integrity and authority.3 In fact, the term Sagan-ization is often used to describe scientists who “become popular enough as an explainer of science to risk the contempt of more ‘serious’ researchers.”4 This is a reference to the widely held notion that the popular Cornell astrophysicist, Carl Sagan, was denied admittance to the National Academy of Sciences because of his publicly televised series, Cosmos.5 Historically, changing socio-cultural patterns and an evolving communication environment have led to renewed attention to scientists’ roles in communicating science outside the ivory tower. The increasing demand for science and technology during World War I put scientists in the public eye more than ever before. Many scientists since then have been under the impression that nationally funded science in the United States could be supported “only if the scientific and nonscientific sectors of American culture were united.”6 This desire for public acceptance of scientific research, especially with respect to emerging technologies with significant social and ethical implications, inspires a “legitimation discourse” of science in media outlets.7 Scientific institutions, as well as some scientists, increasingly orient themselves toward the media. Simultaneously, media are increasingly attentive to scientific research.8 Major scientific institutions and funding agencies also require public communication of science and technology (PCST) components in their funded research.9 The Internet has fundamentally changed our modern media environment and audiences’ media consumption habits.10 The volume of content about science and technology in traditional news outlets has ebbed due to significant declines in readership and subscriptions. In turn, these declines have forced media corporations to decrease the number of journalists who specialize in communicating scientific issues.11 In 1989, there were ninety-five weekly science sections in newspapers in the United States. However, by 2005, fewer than a third of these remained, and that number plunged to nineteen in 2012.12 In light of these changes, the boundaries of communication that exist between scientists, journalists, and public audiences become more blurred. The public relies on various media across both traditional and online platforms for science news and information,13 and almost half of the public turns to online sources to follow developments in scientific fields.14 This poses new opportunities for scientists to play an active role in communicating directly with various publics. However, the question still remains whether public communication efforts by scientists yield any rewards. Researchers have yet to investigate empirically and agree on the impact of communicating one’s work in various media, particularly online media, on scholars’ advancement within the ivory tower. Our study fills this gap in the literature by exploring whether public outreach via traditional and online media can boost scholars’ academic careers. Specifically, we attempt to address whether new media can amplify the effect of traditional public outreach on scholars’ scientific impact.15

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 774 Journalism & Mass Communication Quarterly 91(4)

Scientific Impact Science is a collective endeavor. The impact of scientific research is defined as the extent to which it can benefit other researchers in generating further discoveries,16 and the cumulative impact of a particular researcher’s scientific output, such as publications, adds up to the researcher’s “scientific impact.”17 Because scientific publications play a central role in systematically documenting research findings and facilitating information exchange between researchers, citation analysis is widely acknowledged as a powerful method for quantifying researchers’ scientific impact in order to evaluate and compare scholars in hiring, funding, and tenure decisions.18 Although scientists may alternatively refer to scholars’ “scientific reputation” (opinions generally held by peers about a scholar), the multi-dimensional nature of this notion generates a mix of explicit (e.g., bibliometric indicators) and nebulous measures (e.g., certain valued qualities such as fair play, integrity, honesty, caring, etc.)19 that can be affected by subjectivity and bias.20 In order to utilize fair, transparent, and quantitative approaches to research evaluation,21 our study focuses on measuring scientific impact instead of scientific reputation. The h-index, proposed by Hirsch,22 is a bibliometric indicator that quantifies the scientific impact of a given researcher and embodies a figure of merit.23 According to Hirsch, a researcher has “index h if h of his or her Np papers have at least h citations each and the other (Np − h) papers have ≤ h citations each,”24 so that a high value in h-index indicates a high scientific impact of the researcher. The convergent validity of the h-index has been confirmed in different research fields25 and is robust against small errors and single peaks (top-cited papers) in the publication list of a researcher.26 Although the h-index is sensitive to many factors and should be used with caution,27 it has been widely accepted in the scientific community28 due to its accessibility in citation databases (e.g., Thomson Reuters Web of Science) and its advantages over other bibliometric measures, such as total citation count, citations per paper, and total paper count.29

Scientific Impact Meets Public “Buzz” In academia, articles that receive more attention from other scholars in terms of citations are generally considered more important and prestigious, and the relative importance of the other articles that cite it also determines its impact. This idea of scientific impact, a type of “academic buzz,” is not unique to scholarly work. For example, the algorithm used by the online search engine Google, PageRank, was originally based on this concept.30 PageRank positions webpages referenced by many other popular sites as more important, and thus higher in the results of a search. In this sense, the algorithm that calculates a webpage’s importance is based on the same logic that evaluates peer- reviewed articles’ importance and scholars’ academic impact, but on a much larger scale.

Scientists’ Interactions with Mass Media and Lay Publics As science journalism has shifted from traditional to online media platforms, scientists are interacting with reporters more frequently and seamlessly.31 Science journalists

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 775

increasingly quote peer-reviewed articles in their stories as a way to gain credibility and readers’ trust.32 Frequent interactions with reporters can increase the visibility and popularity of a scientist and his or her work, as they are more likely to be referenced in journalistic narratives. Such narratives can boost the information transmission from scientific literature to the scientific community and, further, to the general public, while gaining more citations by fellow scholars.33 On the individual level, scholars who have frequent media contact tend to be more academically active.34 Yet, contrary to the above findings, the perception of “Sagan-ization” is still prevalent in academia. While it seems reasonable to assume that concordant relationships between scientists and journalists could translate into greater impacts within academia for researchers, the possibility of critical reactions from peers may weaken the potential rewards of such interactions.35 Due to a paucity of empirical evidence exploring whether interactions with journalists impact scholars’ careers, we propose the following research question:

RQ1: With other factors held constant, do scientists’ interactions with reporters affect scientific impact?

In addition to media interactions, scientists’ efforts to engage lay publics and popu- larize their research can be wide-ranging, and may include public speeches, school presentations, and collaborations with other non-academic associations. Contrary to the perception of “Sagan-ization,” scientists who are active in disseminating their work to lay audiences also perform better than average academically. A study of French scientists, for example, showed that scientists who engaged in more dissemination activities for non-specialized audiences published more peer-reviewed articles and were cited more times per year over their research career than the less engaged scientists.36 Given that the existing data demonstrate a positive relationship between outreach activities and academic performance in terms of scientific impact, we put forth the following hypothesis:

H1: With other factors held constant, scientists’ interactions with non-scientists are positively related to scientific impact.

Science Blogging Blogs are a Web 2.0-type tool that have increasingly become a source for the public to get information about scientific developments37 and an open space for scientists from different disciplines to exchange knowledge and evaluate other scientific research.38 Currently, over 26,000 blog entries have been posted about peer-reviewed research on various science subjects on the Research Blogging platform (http://www.research- blogging.org).39 As opposed to scientists who publicize their work by talking with journalists, scientists who blog about their research have more individual autonomy over how their scientific developments are communicated to the public. Scientists may blog to circumvent traditional media outlets to highlight their own recently published work and communicate with peers.40 Some scientists also rely on scientific blogs to

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 776 Journalism & Mass Communication Quarterly 91(4) survey their academic environments.41 In addition, scientific blogs may bring issues that are not yet popularized to the attention of the mainstream media.42 Consistently and regularly updating a blog with scientific achievements may, therefore, help scientists increase the visibility of their published research, including among their peers. In addition to being an open marketplace for scientific exchange, online communication tools, such as blogs, allow researchers to expand their professional networks through online activities. Evidence suggests that scholars who connect through online environments, including blogs, also collaborate on projects offline.43 Although studies on scientists’ online behaviors are relatively few, those that exist provide support for a positive association between blogging and popularity of a scientist’s research. On the basis of this reasoning, we put forth the following hypothesis:

H2: With other factors held constant, scientists’ blogging about science is positively related to scientific impact.

Twitter Activity Twitter is the United States’ second largest social networking platform, with 16% of all Internet users having Twitter accounts.44 The platform provides unique opportunities for scientists to post “tweets,” user-generated content with a limit of 140 words, and offers “various degrees of social presence/media richness.”45 In contrast to other online social networking sites (such as Facebook) that control information sharing only with approved “friends,” public Twitter posts allow live dialogues visible to anyone unless the user opts to use a private setting. Given the sheer size of Twitter users, the open access, and the relative ease of composing tweets, information shared on Twitter by certain opinion leaders, including some prestigious science writers, can immediately reach a large number of audiences. For example, every tweet from Carl Zimmer may be seen by over 146,000 followers and potentially greater audiences when the post is retweeted by these followers. In a survey of higher education professionals, researchers found academic use of Twitter has increased among scholars, with 35.2% of surveyed college faculty members using Twitter in 2010, compared with only 30.7% in 2009.46 In particular, scholars on Twitter were found to be discussing academic conferences and articles.47 Such discussions on Twitter are often legitimate, interactive, wide-ranging, and cross-disciplinary conversations that are reflective of academic impact.48 Being cited or mentioned on Twitter could be a new sign of one’s academic impact.49 Eysenbach mined tweets that mentioned published articles in a medical journal (i.e., tweets with reference to the title and URL of journal articles) and found that journal articles mentioned on Twitter were more likely to be frequently cited by other scholars.50 The current study takes a similar approach to examining mentions of scientists’ research on Twitter, which are measured as tweets that include scientists’ names and their research with links to information on other websites. On the basis of previous findings, we presume that if scientists are mentioned on Twitter, their research may be more visible, which influences the underlying impact of their work:

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 777

H3: With other factors held constant, being mentioned on Twitter is positively related to scientific impact.

Building Buzz in New Media Environments Contemporary media environments have important implications on how scientists monitor scientific developments and communicate about their research. For example, American neuroscientists now rely on an array of cross-media channels, including traditional journalistic outlets (e.g., newspapers, magazines, radio, and television), new media (e.g., blogs), and interpersonal social networks to keep abreast of new research.51 In addition, scientists’ traditional forms of public outreach, usually interacting with journalists who cover their research in the mass media, can be further dis- seminated through Web 2.0-type tools. Therefore, in addition to traditional communication efforts undertaken by researchers, it is reasonable to assume that the use of multiple online channels can amplify the effect of other forms of outreach on researchers’ scientific impact. On the other hand, if there is a “Sagan-ization” effect, some scholars may argue that scientists who are too engaged (e.g., tweet too often) will suppress the impact of other forms of outreach (e.g., interactions with reporters). Researchers have yet to provide empirical evidence that such interactions between various forms of outreach exist. We therefore put forth the following research question:

RQ2: Do different forms of communication behaviors (i.e., interactions with journalists or other non-scientists, science blogging, and being mentioned on Twitter) moderate each other’s effect on scientific impact?

Methods Sample Our sample consists of only the most highly cited U.S. scientists within the field of nanotechnology. Nanotechnology is an emerging and complex field that encompasses a broad area of expertise, drawing from the fields of chemistry, materials science, physics, engineering, biology, and others. Its inventions are integrated with modern biology, the digital revolution, and cognitive sciences.52 We focus on nano-scientists for two reasons. One, elite experts in one discipline may not have an equivalent status in another discipline. By focusing on scientists working in this multidisciplinary field, we can remove the effects of name recognition, which otherwise can be a confounding factor that influences citation patterns and h-indices. Two, the multidisciplinary nature of nanotechnology makes nano-scientists especially pertinent and representative of scientists who work in an evolving scientific community in which the distinctions between disciplines are blurring and research endeavors require interdependence among disciplines.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 778 Journalism & Mass Communication Quarterly 91(4)

We sampled authors of the most cited publications indexed in the Thomson Reuters Web of Science database in 2008 and 2009 in order to minimize the potential confounding effects of seniority on the h-index.53 In order to rigorously establish which publications were actually within the multidisciplinary field of nanotechnology, we relied on a database that indexed a total of 189,014 nanotechnology-related journal articles published in the two-year period of 2008-2009. This database of nanotechnology publications was built upon a set of bibliometric search terms that define the domain of nanotechnology-related publications.54 Using this database, we identified a sample of 1,405 U.S.-affiliated authors of the most highly cited nanotechnology publications, each of whom was cited no fewer than thirty-nine times in the two-year period.

Data Collection Data for the study were collected in two parts. First, a nationally representative survey of leading U.S. nano-scientists was collected by mail. The survey was fielded in four waves between June and September 2011, following Dillman, Smyth, and Christian’s tailored design method.55 A postcard announced the survey to nano-scientists and was followed by an initial mailing of the survey. Next, a postcard reminder was mailed to non-respondents three to four days after the initial mailing, followed by the second mailing of the survey after three to four weeks. The mail survey yielded 444 completed questionnaires, with a final response rate of 31.6%, following American Association for Public Opinion Research’s method of response rate 3.56 Such a response rate is not uncommon in the social sciences, particularly in elite or expert surveys.57 We surveyed respondents about their perceived interactions with journalists and lay publics and the frequency with which they blogged about scientific research. The survey also focused on non-communication issues, such as respondents’ perceptions about ethical, social, and policy implications of nanotechnology, which reduced the likelihood of a nonresponse bias. In order to examine a link between scientists’ public communication behaviors and indicators of scientific impact, we allowed respondents’ h-indices to accumulate over a period of fifteen to eighteen months following our survey and thus collected the second part of our data in December 2012. We then gathered h-indices of all respondents from the Thomson Reuters Web of Science database and recorded cases in which their research was mentioned in tweets. Information from respondents’ curricula vitae, obtained online from the institutions with which they were affiliated, was used to refine our h-index search. Our analysis focused only on scientists in tenure-track faculty positions, so scientists associated with private industry and in federal government positions (e.g., U.S. Department of Agriculture or U.S. Environmental Protection Agency) were excluded due to the lack of accessible curricula vitae. Our final sample was 241 U.S. nano-scientists.

Measures Dependent variable. We used the h-index (M = 37.1, SD = 23.7) as a measure of a researcher’s scientific impact.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 779

Independent variables. Our questions about respondents’ communication behaviors followed the measures of self-reported media use and face-to-face interactions used in previous communication research.58 For the sake of simplicity, we use terms such as public science communication and communication behaviors in the remainder of this article to refer to self-reported communication activities. To obtain a measure of scientists’ interactions with reporters, we asked respondents how often they spoke to reporters about their research findings, based on a 4-point scale (1 = never, 4 = often) (M = 2.6, SD = 0.9). Interactions with other non-scientists was measured by asking respondents how often they talked with non-scientists about their research findings, coded on the same scale (M = 3.1, SD = 0.7). Science blogging was gauged by asking respondents how frequently they wrote a blog about science, using the same 4-point scale (M = 1.3, SD = 0.6). We defined mentions on Twitter as tweets from any Twitter user that referenced the respondent’s name and research with hyperlinks to detailed information. Due to the low number of tweets that mentioned respondents’ research, we chose a dichotomous variable to indicate whether the participant’s own research had been mentioned on Twitter (14.1% were mentioned on Twitter).

Control variables. We controlled for participants’ gender (85.9% male), scientific age (the number of years since his or her first publication; M = 21.2, SD = 10.7), tenure (whether they were tenured faculty members; 73.8% tenured), and the disciplinary field in which they received their doctoral degree (33.1% chemistry, 17.6% engineering, 17.2% physics, 14.2% materials sciences, and 17.9% biology and other sciences) because of sensitivity of the h-index to each of these factors.59 The disciplinary variables were entered in the regression model as a series of dummy variables, with biology and other sciences as the reference group.

Data Analysis We tested our hypotheses and research questions using a hierarchical ordinary least squares (OLS) regression model. The variables were entered in blocks according to their assumed causal order. In the model, the blocks were ordered as follows:

1. Demographics and professional status (gender, scientific age, tenure) 2. Disciplinary field (chemistry, engineering, physics, materials sciences) 3. Public science communication (interactions with reporters, interactions with non-scientists, science blogging, mentioned on Twitter) 4. Two-way interactions

The final block included interaction terms that were created by multiplying standardized versions of the variables to minimize multicollinearity between the interaction terms and their components in the model.60

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 780 Journalism & Mass Communication Quarterly 91(4)

Table 1. Unique Variance Explained by Each Block in the OLS Regression Model Predicting h-Index (N = 241).

R2 (%) Demographics and professional status 35.1*** Disciplinary field 0.8 Science communication 6.5*** Shared variance (%) 17.6*** Total variance (%) 60.0***

Note. OLS = ordinary least squares. ***p ≤ .001.

Results Overall, our model fit the data well, with variables included accounting for 60% of the variance in h-index. Most of the variance was accounted for by demographics and professional status (35.1%), while public science communication variables accounted for 6.5% of the variance in h-index (Table 1). Scientific age (β = .54, p ≤ .001) and tenure (β = .14, p ≤ .05) were both positively related to scientific impact. Senior researchers, or those who had published their first paper earlier relative to others in the sample, had higher h-indices. Tenured scholars also had higher h-indices than those who were not tenured. Our first research question (RQ1) was related to scientists’ communication efforts through more traditional means, measured by their interactions with reporters. We found a positive relationship between interactions with reporters and h-indices (β = .22, p ≤ .001), implying that scholars who had more interactions with reporters had greater scientific impact than those who had fewer interactions with reporters. Neither interactions with other non-scientists nor science blogging was significantly related to h-indices (Table 2). Thus, H1 and H2 were not supported. As hypothesized in H3, scientists whose research was mentioned on Twitter had significantly higher h-indices (β = .13, p ≤ .01) than their peers whose research was not mentioned on Twitter (Table 2). In response to our second research question (RQ2) on the moderating effects of different forms of public communication on the h-index, we found two significant interactions (Table 2). The interactive effect between scientists’ interactions with reporters and being mentioned on Twitter was positive (β = .14, p ≤ .05). Interactions with reporters had a significantly higher impact on the h-index for those scientists who were also mentioned on Twitter than for those who were not (Figure 1). Being mentioned on Twitter also further ampli- fied the effect of interactions with other non-scientists on the h-index (β = .11, p ≤ .05). In other words, the h-indices of scientists who interacted with other non-scientists were higher if they were also mentioned on Twitter, compared with scholars who were not (Figure 2).

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 781

Table 2. OLS Regression Model Predicting h-Index (N = 241).

Zero-order β Block 1: Demographics and professional status Gender (female = 1) −.08 .02 Scientific age .70*** .54*** Tenure (tenured = 1) .54*** .14* Incremental R2 (%) 51.2*** Block 2: Disciplinary field Chemistry .04 −.02 Engineering −.13 −.10 Physics .04 −.05 Material Science −.06 −.07 Incremental R2 (%) 0.2 Block 3: Science communication Interactions with reporters .34*** .22*** Interactions with other non-scientists .18** .02 Science blogging −.03 −.06 Mentioned on Twitter (mentioned = 1) .23*** .13** Incremental R2 (%) 6.5*** Block 4: Two-way interactions Interactions with reporters × Interactions — .08 with non-scientists Interactions with reporters × Science — .01 blogging Interactions with reporters × Mentioned — .14** on Twitter Interactions with non-scientists × Science — .03 blogging Interactions with non-scientists × — .11* Mentioned on Twitter Science blogging × Mentioned on Twitter — .04 Total R2 (%) 60.0***

Note. Cell entries are final standardized regression coefficients for blocks 1, 2, and 3 and before-entry standardized regression coefficients for block 4. OLS = ordinary least squares. *p ≤ .05. **p ≤ .01. ***p ≤ .001.

Discussion In this study, we surveyed the most highly cited U.S. nano-scientists and explored the effects of scientists’ public communication behaviors via traditional and new media on their scientific impact as measured by the h-index. The current study provides the first comprehensive empirical evidence that outreach activities, such as interactions with reporters and being mentioned on Twitter, can assist a scientist’s career by promoting his or her scientific impact. More importantly, online buzz (e.g., being mentioned on

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 782 Journalism & Mass Communication Quarterly 91(4)

Figure 1. Interactive effect between frequency of interaction with reporters and being mentioned on Twitter on h-index. Note. Scale on Y-axis only is partially displayed.

Twitter) further amplifies the impact of communicating science through traditional outlets on the scholar’s scientific impact. Neither science blogging nor interacting with non-scientists had any significant effect on scientific impact, which could be explained by the inherent complexity of each of the two forms of activities. The readership of science blogs may vary greatly from one post to another, and similarly, “interaction with non-scientists” can be wide-ranging (such as talking to family members and collaborating with industry professionals). If this is indeed the case, these two forms of communication activities are not consistently related to scientists’ academic impact. Before elaborating on the implications of our findings, it is important to discuss several limitations of the current work. First, the h-index is not a perfect indicator of scientific impact and should be interpreted with caution. In general, the recognized problems with the h-index include its potential to hamper the measured impact of scientists who have published a small number of papers and its bias across disciplines that have different inherent citation patterns.61 Our sample design (specifically focusing only on the most highly cited authors) and controls in the regression model (e.g., scientific age and disciplines) was constructed to minimize such biases. Despite the potential limitations, the h-index is able to give relatively reliable information about the scientific impact of a given researcher, and is recognized as an improvement in

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 783

Figure 2. Interactive effect between frequency of interaction with other non-scientists and being mentioned on Twitter on h-index. Note. Scale on Y-axis only is partially displayed.

comparison with many other available indices of scientific impact, such as total citation count, citations per paper, and total paper count.62 As noted previously, the h-index has a linear and positive relationship with a scholar’s scientific age.63 The significant relationship we found between the h-index and scientific age demonstrates the construct validity of the h-index measure used in our study. A second concern is related to our relatively small sample size and the generalizability of our findings. The present sample design includes only university-based scientists, which does not allow us to compare subgroups of scientists based on their affiliations with industry or other non-academic institutions. Despite the relatively small sample size, our sample and data are unique and valuable in that we were successful in collecting data from hard-to-reach experts. To our knowledge, there are no previous studies that evaluate the effect of communication efforts on scientific impact that match our research design. More importantly, our sample design is also a strength because it attenuates potential concerns about endogeneity by limiting our analyses to a group of already highly visible scientists. The issue of endogeneity, if not addressed appropriately, could confound our evaluation of the effects of various communication behaviors on one’s scientific impact. Some confounding factors, such as scientists’

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 784 Journalism & Mass Communication Quarterly 91(4) educational institutions, professional status, and h-indices scientists have accumulated in previous years, could be highly correlated with one’s communication behaviors and h-indices in the following years. In other words, these scientists were likely to come from elite educational institutions, have published a number of highly impactful papers, and were therefore covered more frequently in the media. In this study, the issue of endogeneity is minimized in three ways. First, as presented above, we focused on a heterogeneous sample of the most highly cited scientists. Second, we collected scientists’ h-indices about one-and-a-half years after surveying their communication behaviors. Third, we controlled for the factors that might be correlated with both scientists’ communication behaviors and their scientific impact, for example, gender and professional status (scientific age, whether the respondent was tenured, and disciplinary field). As a result, we observed a significant and positive association between active communication behaviors and the h-index after their communication behaviors. It is reasonable to assume that the strength of the observed associations would increase if we adopted a longer time period to allow h-indices to accumulate following various communication behaviors. Furthermore, it is important to take into account the nature of our operationalization of activities on Twitter. Ideally, we would like to have included continuous measures of both active and passive Twitter activities, that is, scientists’ tweeting research updates to their followers, as well as being mentioned in others’ tweets. However, too few scientists were active Twitter users to include either active Twitter use or a continuous measure (as opposed to our dichotomous indicator) of Twitter mentions in our study. The limited number of respondents using Twitter was unsurprising, given that tweeting about scientific research is a relatively recent phenomenon within academia, and currently 16% of the general population are on Twitter.64 Nonetheless, Twitter should still be viewed as a critical platform for science information exchange and public science communication, due in part to the sheer volume of science-related posts on Twitter; for example, there were over 495,000 nanotechnology-related opinions shared on Twitter over just one year between September 1, 2010, and August 31, 2011.65 More importantly, the number of Twitter users has grown at an enormous rate, with the proportion of Internet users who are on Twitter doubling since the end of 2010.66 On these bases, the influence of tweeted nano-related information can be important. As noted above, we did not focus on how each single tweet affects the impact of single news stories in traditional media or how the salience of scientific issues transfers from one medium to another. Instead, we aimed to answer a more important question: the overall effect of Twitter activities on the link between traditional communication efforts and scientific impact. Mindful of the limitations of the current study, the primary finding that the professional status of nano-scientists may benefit from mass media interactions contrasts with the conclusions made within prior science communication literature.67 This is particularly important when mass communication is undergoing significant transfor- mation and science is expanding its role in society. Nowadays, collaborations between journalists and scientists are increasingly frequent,68 despite the shrinking science- related news hole in the mass media. Almost all the two thousand members of the

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 785

National Association of Science Writers are freelancers who depend on working relationships with individual scientists as information sources.69 The positive link between scientist–journalist interactions and scientific impact may, therefore, encourage scientists’ engagement with science communication through legacy media, which will ultimately serve lay audiences as well. Another significant finding of this study indicates that increasing use of online media compared with legacy media may be impacting contemporary science communication. In Web 2.0, the boundaries that separated scientists, journalists, and the public may be blurring. Our findings suggest there is value in “building buzz” by utilizing social media as well as legacy mass communication channels to enrich information exchanges between the scientific community and public audiences. Many scholars have suggested that social media are supplementing rather than supplanting conventional channels, such as newspapers and television, for scientific information.70 For the moment, this may be true. In particular, social media can augment the impact of more conventional forms of public communication as demonstrated in this study. Yet social media may also present the scholarly community with new challenges related to traditional metrics of success. In academic circles, book blurbs from well- known scholars or book reviews by prestigious media outlets are generally considered more impactful than those from less well-known entities. A similar logic applies to social media. If a scientist’s work is tweeted by prominent science reporters (such as Andrew Revkin, who has more than 50,000 followers on Twitter), scientists (such as Neil deGrasse Tyson with more than 1.5 million followers on Twitter), or science media outlets (such as Science Friday with over 446,000 followers on Twitter), it is likely to attract more attention and have a larger impact even within academic circles, than a study that was only published in a peer-reviewed academic outlet (even for elite outlets, such as Nature and Science with impact factors of 36.28 and 31.2, respectively). The rewards for public communication efforts on social media may eventually force academics to think more carefully about mapping academic impact in a world of sites, such as Google Scholar and ResearchGate.com, which combine social media metrics with indicators of scholarly productivity to measure the broader impact of academic work. Indeed, some scholars have recently called for social media to be used to supplement traditional approaches to measuring academic impact.71 Nonetheless, it is noteworthy that current online social media environments may have potential pitfalls for science communicators, and mass communication at large. Open and interactive dialogues inherent to Web 2.0 tools like Twitter and Facebook enable audiences to repurpose and translate scientists’ research findings using their own interpretations and debate them on social media.72 Thus, social networks can also help spread potential misinterpretations of scientific findings quickly among large audiences. For example, some scholars have raised concerns that readers’ uncivil online comments following scientific information on social media can polarize perceptions of risks associated with a technology73 and even bias perceptions of source and message credibility.74 Future research could conduct a more fine-grained exploration of scientists’ public outreach efforts with sophisticated data collection. For example, a study can include

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 786 Journalism & Mass Communication Quarterly 91(4) more time points to collect data of different patterns of communication behaviors associated with both traditional media and new media to explore how each behavior sequentially affects scholars’ scientific impact. The variation in institutional and disciplinary culture should also be given consideration. A larger sample size would allow for more careful examination of the specific attributes of scientists across disciplines and affiliations. Comparing the impacts of outreach activities for scientists with different affiliations (such as an industry-based, government-based, versus university-based comparison) and from other research disciplines could yield distinct findings. In addition, future scholarship should use precise measurements of scientists’ communication behaviors. In particular, measures with a reference point and actual frequencies (e.g., times per month) would capture actual public communication behaviors by scientists. Finally, we encourage attempts to obtain continuous variables of both active and passive activities on social media, such as frequencies of reposting one’s own research, posting comments on others’ research, and mentions of one’s research. These approaches could yield valuable results about the impact of various increasingly popular social media (such as Facebook, Twitter, and Google Plus) on scientists’ careers.

Authors’ Note Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This material is based upon work supported by grants from the National Science Foundation to the University of Wisconsin-Madison Nanoscale Science and Engineering Center in Templated Synthesis and Assembly at the Nanoscale (Grant SES- DMR-0832760) and the Center for Nanotechnology in Society at Arizona State University (Grant SES-0937591).

Notes 1. “Survey of Factors Affecting Science Communication by Scientists and Engineers,” TheRoyal Society, 2006, https://royalsociety.org/~/media/Royal_Society_Content/policy/ publications/2006/1111111395.pdf(accessed May 10, 2013). 2. Michael Shortland and Jane Gregory, Communicating Science: A Handbook (NY: Longman, 1991). 3. Sharon Dunwoody and Michael Ryan, “Scientific Barriers to the Popularization of Science in the Mass Media,” Journal of Communication 35 (1, 1985): 26-42; Felicity Mellor, “Negotiating Uncertainty: Asteroids, Risk and the Media,” Public Understanding of Science 19 (1, 2010): 16-33.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 787

4. Donald Kennedy, “Science and the Media,” in Science and the Media, ed. Donald Kennedy and Geneva Overholser (Cambridge, MA: American Academy of Arts and Sciences, 2010), 1-10. 5. Cornelia Dean, Am I Making My Self Clear? A Scientist’s Guide to Talking to the Public (Cambridge, MA: Harvard University Press, 2009). 6. Ronald C. Tobey, The American Ideology of National Science, 1919-1930 (Pittsburgh, PA: University of Pittsburgh Press, 1971), 167. 7. Peter Weingart, Die Wissenschaft Der Öffentlichkeit [The Science of Public Opinion] (Weilerswist, Germany: Velbrück Verlag, 2005). 8. Simone Rödder, “Reassessing the Concept of a Medialization of Science: A Story from the ‘Book of Life,’” Public Understanding of Science 18 (4, 2009): 452-63. 9. Rödder, “Reassessing the Concept”; Gillian Pearson, “The Participation of Scientists in Public Understanding of Science Activities: The Policy and Practice of the U.K. Research Councils,” Public Understanding of Science 10 (1, 2001): 121-37. 10. Dominique Brossard and Dietram A. Scheufele, “Science, New Media, and the Public,” Science 339 (January 4, 2013): 40-41; Ashley A. Anderson, Dominique Brossard, and Dietram A. Scheufele, “The Changing Information Environment for Nanotechnology: Online Audiences and Content,” Journal of Nanoparticle Research 12 (February 7, 2010): 1083-94. 11. Anthony Dudo, Sharon Dunwoody, and Dietram A. Scheufele, “The Emergence of Nano News: Tracking Thematic Trends and Changes in Media Coverage of Nanotechnology,” Journalism & Mass Communication Quarterly 88 (spring 2011): 55-75. 12. Sara Morrison, “Hard Numbers: Weird Science,” Columbia Journalism Review, January 2, 2013, http://www.cjr.org/currents/hard_numbers_jf2013.php (accessed February 18, 2013). 13. Leona Yi-Fan Su, Ashely A. Anderson, Dominique Brossard, Dietram A. Scheufele, and Mike A. Xenos, “Audience Tectonics: Implications of Changing News Environments for Public Understanding of Science” (paper presented at the annual meeting for the International Public Communication of Science and Technology, Florence, Italy, April 18-20, 2012). 14. National Science Board, Science and Engineering Indicators 2012 (Arlington, VA: National Science Foundation, 2012). 15. Throughout this article, we use “academic impact” and “scientific impact” as analogous terms. 16. Nigel Shadbolt, Tim Brody, Les Carr, and Stevan Harnad, “The Open Research Web: A Preview of the Optimal and the Inevitable,” in Open Access: Key Strategic, Technical and Economic Aspects, ed. Neil Jacobs (Oxford, UK: Chandos, 2006), 195-208. 17. Jorge E. Hirsch, “An Index to Quantify an Individual’s Scientific Research Output,” Proceedings of the National Academy of Sciences of the United States of America 102 (2005): 16569-72. 18. Christine L. Borgman and Jonathan Furner, “Scholarly Communication and Bibliometrics,” Annual Review of Information Science and Technology 36 (2002): 2-72; Hirsch, “An Index to Quantify an Individual’s Scientific Research Output”; Julian Warner, “A Critical Review of the Application of Citation Studies to the Research Assessment Exercises,” Journal of Information Science 26 (6, 2000): 453-60; Peter Weingart, “Impact of Bibliometrics upon the Science System: Inadvertent Consequences?” Scientometrics 62 (1, 2005): 117-31. 19. Philip E. Bourne and Virginia Barbour, “Ten Simple Rules for Building and Maintaining a Scientific Reputation,” PLoS Computational Biology 7 (6, 2011): e1002108, accessed

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 788 Journalism & Mass Communication Quarterly 91(4)

February 10, 2013, doi:10.1371/journal.pcbi.1002108; Cristhian Parra, Fabio Casati, Florian Daniel, Maurizio Marchese, Luca Cernuzzi, Marlon Dumas, Peep Kungas, Luciano García-Bañuelos, and Karina Kisselite, “Investigating the Nature of Scientific Reputation” (paper presented at the annual meeting for the International Society for Scientometrics and Informetrics, Durban, South Africa, July 4-7, 2011). 20. Parra et al., “Investigating the Nature of Scientific Reputation.” 21. Philip Ball, “Achievement Index Climbs the Ranks,” Nature 448 (August 2007): 737. 22. Hirsch, “An Index to Quantify an Individual’s Scientific Research Output.” 23. Jorge E. Hirsch, “Does the H Index Have Predictive Power?” Proceedings of the National Academy of Sciences of the United States of America 104 (2007): 19193-98; Anthony F. J. van Raan, “Comparison of the Hirsch-Index with Standard Bibliometric Indicators and with Peer Judgment for 147 Chemistry Research Groups,” Scientometrics 67 (3, 2006): 491-502. 24. Hirsch, “An Index to Quantify an Individual’s Scientific Research Output,” 16569. 25. Lutz Bornmann and Hans-Dieter Daniel, “Selecting Scientific Excellence through Committee Peer Review—A Citation Analysis of Publications Previously Published to Approval or Rejection of Post-Doctoral Research Fellowship Applicants,” Scientometrics 68 (3, 2006): 427-40; Lutz Bornmann and Hans-Dieter Daniel, “Convergent Validation of Peer Review Decisions Using the H Index: Extent of and Reasons for Type I and Type II Errors,” Journal of Informetrics 1 (3, 2007): 204-13; Blaise Cronin and Lokman Meho, “Using the H-Index to Rank Influential Information Scientists,” Journal of the American Society for Information Science and Technology 57 (9, 2006): 1275-78; Clint D. Kelly and Michael D. Jennions, “The H Index and Career Assessment by Numbers,” Trends in Ecology & Evolution 21 (4, 2006): 167-70; van Raan, “Comparison of the Hirsch-Index.” 26. Yuxian Liu and Ronald Rousseau, “Properties of Hirsch-Type Indices: The Case of Library Classification Categories,” Scientometrics 79 (2, 2009): 235-48. 27. Monica Gaughan and Branco Ponomariov, “Faculty Publication Productivity, Collaboration, and Grants Velocity: Using Curricula Vitae to Compare Center-Affiliated and Unaffiliated Scientists,” Research Evaluation 17 (2, 2008): 103-10; Hirsch, “An Index to Quantify an Individual’s Scientific Research Output”; Kelly and Jennions, “The H Index and Career Assessment by Numbers”; Jean King, “A Review of Bibliometric and Other Science Indicators and Their Role in Research Evaluation,” Journal of Information Science 13 (5, 1987): 261-76; Hanna Kokko and William J. Sutherland, “What Do Impact Factors Tell Us?” Trends in Ecology & Evolution 14 (10, 1999): 382-84. 28. Philip Ball, “Index Aims for Fair Ranking of Scientists,” Nature 436 (7053, 2005): 900; Lutz Bornmann and Hans-Dieter Daniel, “Does the H-Index for Ranking of Scientists Really Work?” Scientometrics 65 (3, 2005): 391-92. 29. Ball, “Achievement Index Climbs the Ranks”; Hirsch, “An Index to Quantify an Individual’s Scientific Research Output”; Hirsch, “Does the H Index Have Predictive Power?” 30. Eli Pariser, The Filter Bubble: What the Internet Is Hiding from You (NY: The Penguin Press, 2011). 31. Sharon Dunwoody, Dominique Brossard, and Anthony Dudo, “Socialization or Rewards? Predicting U.S. Scientist-Media Interactions,” Journalism & Mass Communication Quarterly 86 (June 2009): 299-314; Hans Peter Peters, Dominique Brossard, Suzanne de Cheveigné, Sharon Dunwoody, Monika Kallfass, Steve Miller, and Shoji Tsuchida, “Science-Media Interface: It’s Time to Reconsider,” Science Communication 30 (2, 2008): 266-76.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 789

32. Vinciane Colson, “Science Blogs as Competing Channels for the Dissemination of Science News,” Journalism 12 (7, 2011): 889-902. 33. Vincent Kiernan, “Diffusion of News about Research,” Science Communication 25 (1, 2003): 3-13; David P. Phillips, Eilliot J. Kanter, Bridget Bednarczyk, and Patricia L. Tastad, “Importance of the Lay Press in the Transmission of Medical Knowledge to the Scientific Community,” New England Journal of Medicine 325 (16, 1991): 1180-83. 34. Pablo Jensen, Jean-Baptiste Rouquier, Pablo Kreimer, and Yves Croissant, “Scientists Who Engage with Society Perform Better Academically,” Science and Public Policy 35 (7, 2008): 527-41. 35. Dunwoody, Brossard, and Dudo, “Socialization or Rewards?” 36. Jensen et al., “Scientists Who Engage with Society.” 37. Julie Jones and Itai Himelboim, “Just a Guy in Pajamas? Framing the Blogs in Mainstream US Newspaper Coverage (1999-2005),” New Media & Society 12 (2, 2010): 271-88. 38. Colson, “Science Blogs.” 39. Sibele Fausto, Fabio A. Machado, Luiz Fernando J. Bento, Atila Iamarino, Tatiana R. Nahas, and David S. Munger, “Research Blogging: Indexing and Registering the Change in Science 2.0,” PLoS ONE 7 (12, 2012): e50109, accessed March 10, 2013, doi:10.1371/ journal.pone.0050109. 40. Colson, “Science Blogs”; Dominique Brossard, “A Brave New World: Challenges and Opportunities for Communicating about Biotechnology in New Information Environments,” in Biotechnologie-Kommunikation: Kontroversen, Analysen, Aktivitäten, ed. Marc-Denis Weitze, Alfred Puehler, Wolfgang M. Heckl, Bernd Mueller-Roeber, Ortwin Renn, Peter Weingart, and Gunther Wess (Heidelberg, Germany: Springer, 2012), 427-45. 41. Joachim Allgaier, Sharon Dunwoody, Dominique Brossard, Yin-Yueh Lo, and Hans Peter Peters, “Journalism and Social Media as Means of Observing the Contexts of Science,” BioScience 63 (4, 2013): 284-87. 42. Michael A. Cacciatore, Ashley A. Anderson, Doo-Hun Choi, Dominique Brossard, Dietram A. Scheufele, Xuan Liang, Peter J. Ladwig, Michael Xenos, and Anthony Dudo, “Coverage of Emerging Technologies: A Comparison between Print and Online Media,” New Media & Society 14 (6, 2012): 1039-59. 43. Colson, “Science Blogs.” 44. “The Demographics of Social Media Users—2012,” Pew Research Center, 2013, http:// www.pewinternet.org/files/old-media//Files/Reports/2013/PIP_SocialMediaUsers.pdf (accessed March 3, 2013). 45. Andreas M. Kaplan and Michael Haenlein, “Users of the World, Unite! The Challenges and Opportunities of Social Media,” Business Horizons 53 (1, 2010): 59-68, 61. 46. “Twitter in Higher Education 2010: Usage Habits and Trends of Today’s College Faculty,” Faculty Focus, 2010, http://www.facultyfocus.com/wp-content/uploads/images/2010-twitter-survey-report.pdf (accessed March 10, 2013). 47. Jeffrey R. Young, “10 High Fliers on Twitter,” Chronicle of Higher Education 55 (31, 2009): A10, http://chronicle.com/article/10-High-Fliers-on-Twitter/16488/ (accessed March 20, 2013). 48. Jason Priem and Kaitlin Light Costello, “How and Why Scholars Cite on Twitter,” Proceedings of the American Society for Information Science and Technology 47 (2010): 1-4. 49. Jason Priem and Bradely H. Hemminger, “Scientometrics 2.0: Toward New Metrics of Scholarly Impact on the Social Web,” First Monday 15 (7, 2010), accessed March 23, 2013, doi:10.5210/fm.v15i7.2874.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 790 Journalism & Mass Communication Quarterly 91(4)

50. Gunther Eysenbach, “Can Tweets Predict Citations? Metrics of Social Impact Based on Twitter and Correlation with Traditional Metrics of Scientific Impact,” Journal of Medical Internet Research 13 (4, 2011): e123, http://www.jmir.org/2011/4/e123/ (accessed March 5, 2013). 51. Allgaier et al., “Journalism and Social Media.” 52. Mihail C. Roco and William Sims Bainbridge, Converging Technologies for Improving Human Performance (Dordrecht, The Netherlands: Kluwer Academic Publishers, 2003). 53. Hirsch, “An Index to Quantify an Individual’s Scientific Research Output”; Kelly and Jennions, “The H Index and Career Assessment.” 54. Alan L. Porter, Jan Youtie, Philip Shapira, and David J. Schoeneck, “Refining Search Terms for Nanotechnology,” Journal of Nanoparticle Research 10 (5, 2008): 715-28. 55. Don A. Dillman, Jolene D. Smyth, and Leah Melani Christian, Internet, Mail, and Mixed- Mode Surveys: The Tailored Design Method (Hoboken, NJ: Wiley, 2008). 56. For a full discussion of response rate, see American Association for Public Opinion Research, Final Dispositions of Case Codes and Outcome Rates for Surveys (Lenexa, KS: American Association for Public Opinion Research, 2011). 57. See, for example, James S. Dietz, Ivan Chompalov, Barry Bozeman, Eliesh O’Neil Lane, and Jongwon Park, “Using the Curriculum Vita to Study the Career Paths of Scientists and Engineers: An Exploratory Assessment,” Scientometrics 49 (3, 2000): 419-42. 58. See, for example, Jennifer Brundidge, “Encountering ‘Difference’ in the Contemporary Public Sphere: The Contribution of the Internet to the Heterogeneity of Political Discussion Networks,” Journal of Communication 60 (4, 2010): 680-700; Joerg Matthes, Kimberly Rios Morrison, and Christian Schemer, “A Spiral of Silence for Some: Attitude Certainty and the Expression of Political Minority Opinions,” Communication Research 37 (6, 2010): 774-800; Dietram A. Scheufele and Bruce V. Lewenstein, “The Public and Nanotechnology: How Citizens Make Sense of Emerging Technologies,” Journal of Nanoparticle Research 7 (6, 2005): 659-67. 59. Gaughan and Ponomariov, “Faculty Publication Productivity”; Hirsch, “An Index to Quantify an Individual’s Scientific Research Output”; Kelly and Jennions, “The H Index and Career Assessment”; King, “A Review of Bibliometric”; Kokko and Sutherland, “What Do Impact Factors Tell Us?”; Steven Stack, “Gender, Children and Research Productivity,” Research in Higher Education 45 (8, 2004): 891-920. 60. Jacob Cohen, Patricia Cohen, Stephen G. West, and Leona S. Aiken, Applied Multiple Regression/Correlation Analysis for the Social Sciences (Hillsdale, NJ: Lawrence Erlbaum, 2003). 61. Leo Egghe, “How to Improve the H-Index,” Scientist 20 (3, 2006): 15; Hirsch, “An Index to Quantify an Individual’s Scientific Research Output”; Thed van Leeuwen, “Testing the Validity of the Hirsch-Index for Research Assessment Purposes,” Research Evaluation 17 (2, 2008): 157-60. 62. Ball, “Index Aims for Fair Ranking of Scientists”; Ball, “Achievement Index Climbs the Ranks”; Hirsch, “An Index to Quantify an Individual’s Scientific Research Output”; Hirsch, “Does the H Index Have Predictive Power?”; Liu and Rousseau, “Properties of Hirsch-Type Indices.” 63. Hirsch, “An Index to Quantify an Individual’s Scientific Research Output.” 64. “The Demographics of Social Media Users—2012.” 65. Kristin K. Runge, Sara K. Yeo, Michael Cacciatore, Dietram A. Scheufele, Dominique Brossard, Michael Xenos, Ashley Anderson, et al., “Tweeting Nano: How Public Discourses

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 Liang et al. 791

about Nanotechnology Develop in Social Media Environments,” Journal of Nanoparticle Research 15 (1, 2013): 1381. 66. “The Demographics of Social Media Users—2012.” 67. See, for example, “Survey of Factors Affecting Science Communication”; Dunwoody and Ryan, “Scientific Barriers.” 68. Dunwoody, Brossard, and Dudo, “Socialization or Rewards?” 69. “About the National Association of Science Writers Inc.,” National Association of Science Writers, 2013, http://www.nasw.org/about-national-association-science-writers-inc (accessed May 10, 2013). 70. See, for example, Allgaier et al., “Journalism and Social Media.” 71. Priem and Hemminger, “Scientometrics 2.0.” 72. Brossard and Scheufele, “Science, New Media, and the Public.” 73. Ashley A. Anderson, Dominique Brossard, Dietram A. Scheufele, Michael A. Xenos, and Peter Ladwig, “The ‘Nasty Effect’: Online Incivility and Risk Perceptions of Emerging Technologies,” Journal of Computer-Mediated Communication 19 (3, 2013): 373-87, accessed March 3, 2013, doi:10.1111/jcc4.12009. 74. Elaine W. J. Ng and Benjamin H. Detenber, “The Impact of Synchronicity and Civility in Online Political Discussions on Perceptions and Intentions to Participate,” Journal of Computer-Mediated Communication 10 (3, 2005): 32.

Downloaded from jmq.sagepub.com at Univ.Di Napoli Federico II - DIPARTIMENTO DI ECONOMIA on December 20, 2014 JMQXXX10.1177/1077699014550091Journalism & Mass Communication QuarterlyWatson 550091research-article2014

Communicating Science and Crises

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 792–810 Assessing Ideological, © 2014 AEJMC Reprints and permissions: Professional, and sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550091 Structural Biases in jmcq.sagepub.com Journalists’ Coverage of the 2010 BP Oil Spill

Brendan R. Watson1

Abstract Previous studies based on aggregate data have not found consistent evidence that journalists’ personal beliefs and attitudes bias their coverage. This study, however, uses individual-level survey data on Gulf Coast journalists’ beliefs and attitudes toward the BP oil spill, matched with a content analysis of respondents’ stories about the disaster, and community structure data. The study examines the effect that journalists’ perceptions of professional norms and the social and economic contexts of the communities in which they work had on their attitudes toward and coverage of the crisis.

Keywords bias, oil spill, community structure, professional roles

Charges persist, particularly in partisan media,1 that reporters’ political ideology biases their coverage of important issues, including the 2010 BP Gulf oil spill, that year’s most-followed domestic news story.2 Some academic studies have found support for the “liberal bias hypothesis,”3 though others have challenged those findings.4 Previous studies, however, have not matched individual journalists’ political beliefs with their coverage. They have relied on aggregate data, which may not accurately reflect how individual journalists’ beliefs shape their coverage. Thus, this study examines whether Gulf Coast journalists’ political beliefs biased their coverage of the BP oil spill by

1University of Minnesota-Twin Cities, Minneapolis, MN, USA Corresponding Author: Brendan R. Watson, School of Journalism & Mass Communication, College of Liberal Arts, University of Minnesota-Twin Cities, Murphy Hall 111, 206 Church St. SE, Minneapolis, MN 55455, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Downloaded from http://www.elearnica.ir Watson 793

matching individual-level survey responses with a content analysis of the stories individual journalists wrote. Previous studies of journalistic bias suggest even these individual-level data are unlikely to produce strong support for the “liberal media hypothesis.”5 This study, however, does not propose that one can predict differences in coverage based on political ideology alone. Journalists, after all, are not completely autonomous of the professional and community contexts in which they are situated.6 The study combines individual-, professional-, and community-level perspectives to suggest that any bias in journalists’ work likely comes from multiple, subtler influences. Specifically, it examines whether contextual factors, including perceived professional norms and the social and economic structure of journalists’ communities, influenced their coverage of the BP oil spill.

Literature Review Journalists’ Beliefs, Attitudes, and Reporter Bias A story can have a slant—that is, favor a particular aspect of a story over another— without being biased.7 For example, political coverage may simply give more positive coverage to the stronger campaign. But coverage is said to be biased when some factor external to the issue or subject being covered—journalists’ political ideology, for example—produces that slanted coverage. Claims of bias are most common when news coverage touches on a controversial issue, such as offshore oil drilling, that exhibits a strong partisan divide.8 Scholars, though, have found little empirical support for the claim that journalists are biased, either toward liberal or conservative causes. D’Alessio and Allen’s meta-analysis, for example, found what they called a “probably insubstantial” pro-liberal bias in network news coverage.9 Lowry and Shidler did find that Bill Clinton received more favorable coverage when he ran against George H. W. Bush in 1992 and Bob Dole in 1996.10 Because, as a whole, journalists lean slightly left of center politically and vote Democrat,11 Lowry and Shidler interpreted this coverage as reflecting journalists’ liberal media bias.12 However, without a direct link between an individual reporter’s political beliefs and the slant of that reporter’s coverage, one cannot determine whether the coverage was slanted due to journalists’ political beliefs, or simply because Clinton was the superior candidate and more deserving of positive coverage.13 This is perhaps why there is more evidence of bias in years Democrats won the presidency (i.e., when Democrats were stronger candidates), but not in years Republicans won.14 A quasi-experiment that matched individual journalists’ political ideologies and their news judgments did find that journalists’ news decisions correlate strongly with their individual political ideologies.15 More conservative journalists were more likely to frame a hypothetical chemical plant controversy in a manner favorable to the chemical industry and against government regulation, whereas liberals favored government regulation. That experiment, however, did not examine actual news coverage,16 nor did a previous survey-based study that found Gulf Coast journalists’ attitudes toward

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 794 Journalism & Mass Communication Quarterly 91(4) the BP oil spill were moderately correlated with their personal political and environmental beliefs.17 Thus, the current research matches individual journalists’ responses to a survey with a content analysis of the stories they wrote to examine whether coverage of the oil spill was affected both by their individual-level political beliefs and attitudes toward the environment.

Professional Roles Different professions, including journalism, have unique cultures, including shared values and norms of how one is to carry out one’s work.18 Based on surveys of journalists’ values and norms, Weaver and colleagues identified four types of journalists: the Interpreters believe journalists’ primary function is to provide analysis of complex problems, Adversarialists believe journalists should be skeptical of public officials and business interests, the Disseminators seek to serve the largest possible audience, and the Populist Mobilizers want to give their readers a voice in the media and encourage them to be actively involved in community affairs.19 Previous studies of journalists’ professional roles have been primarily descriptive in nature.20 Watson, however, found that journalists’ affiliation with the Populist Mobilizer role positively predicted Gulf Coast journalists’ support for additional regulation of the oil industry following the BP oil spill (r = .24).21 That study, however, did not examine how journalists’ attitudes shaped their coverage.

Community Structure To build and maintain information sources and audience for their coverage, journalists must also be aware of and responsive to the larger social structure in which they work. In smaller, more homogeneous (i.e., less structurally pluralistic) communities, power is more likely to be concentrated among a small group of elite actors, and local media are more reticent to cover conflicts, including environmental problems.22 But in more structurally pluralistic communities, society is more reliant on the media to communicate among larger, more dispersed, and more diverse social groups. In these pluralistic communities, media cover conflict more openly. Studies have found that structural pluralism not only affects the amount of coverage environmental problems receive, but also how those issues are framed.23 Griffin, Dunwoody, and Gehrmann, for example, found that newspapers in more structurally pluralistic communities were more likely to use thematic frames in their coverage of local industrial contamination.24 Griffin and Dunwoody, however, found that economic reliance on manufacturing—measured as the percentage of the local workforce employed in manufacturing— was a stronger predictor than structural pluralism of how newspapers framed contamination; newspapers in more economically reliant communities were less likely to link industrial contamination to threats to human health.25 Watson also found structural pluralism negatively predicted Gulf Coast journalists’ attitudes toward the oil

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 795

industry following the BP oil spill, but that local communities’ economic reliance on the oil industry was a stronger predictor of journalists’ attitudes.26 These studies have not tested how community structure variables enter or influence the news-gathering process. Watson suggested that journalists are socialized into the local social structure, which, along with their personal beliefs, shapes their attitudes toward issues in the news, which in turn affect their coverage of those issues.27 Watson’s study, however, examined only journalists’ attitudes, not their coverage.

Journalists’ Coverage of the Oil Spill Previous studies of journalistic bias have focused primarily on whether the tone of a story was positive, negative, or neutral.28 This study focuses more broadly on whether coverage was “critical,” examining the extent to which coverage focused on BP’s responsibility for the oil spill, journalists’ use of thematic frames, and their use of unofficial sources. Kensicki’s content analysis found that while newspapers focused on industry as the cause of pollution, government was overwhelmingly portrayed as being responsible for addressing pollution’s consequences.29 She concluded that this pattern supports business as usual, shifting attention away from the changes needed to address the root causes of pollution (i.e., forcing industry to clean up its act). How a story is framed can also shift how the audience assigns responsibility for a given problem. According to Iyengar, episodic stories—stories framed as isolated incidents—result in responsibility being assigned to individual actors.30 On the other hand, thematic frames, which explore larger causes and consequences underlying a particular issue—for example, putting the BP oil spill in context of the government’s record of inspecting offshore oil drilling operations—are more likely to cause audiences to attribute responsibility for social problems to systemic failures. Exposure to thematic framing of climate change’s effects has also been found to increase support for additional regulations on greenhouse gas-emitting industries.31 Entman and Rojecki also found that stories that use unofficial sources are more likely to raise critical questions.32 Their content analysis of the nuclear freeze movement found that two-thirds of those stories that raised concerns over nuclear weapons attributed those concerns to unofficial sources. Unofficial sources, however, were quoted in only 12% of stories about the nuclear freeze movement. In a content analysis of newspaper coverage of the 1989 Exxon Valdez oil spill, Smith also found that official sources made up 60% of journalists’ sources, and that these government and oil industry sources were most likely to say the crisis had been overblown and to rate Exxon’s response to the crisis favorably.33 While these variables help describe the degree of criticalness present in journalists’ coverage, previous studies have also shown that these variables are affected by the social structure in which journalists work. Journalists in less structurally pluralistic communities are less likely to use unofficial sources34 or thematic frames35 in their coverage. Journalists who work in less pluralistic and more economically dependent

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 796 Journalism & Mass Communication Quarterly 91(4)

Figure 1. Path model: Individual and structural biases in Gulf Coast newspaper journalists’ coverage of the BP oil spill. communities are also less likely to hold a polluting industry responsible for an environmental problem.36

Hypotheses Previous studies have not found strong, consistent evidence of ideological bias in journalists’ coverage.37 The aggregate-level data these studies relied upon, however, fail to link adequately individual journalists’ beliefs and attitudes with their coverage. This study links individual-level survey data on journalists’ attitudes toward the oil industry with a content analysis of these individual journalists’ stories. This study retests the bias hypothesis using a path model, a simplified version of which is shown in Figure 1, which links individual-level survey data with a content analysis of those journalists’ coverage of the BP oil spill.

H1: Journalists’ positive attitudes toward the oil industry will negatively predict their critical coverage of the oil industry; journalists with more positive attitudes will be less likely to write negative stories (H1a), focus on BP’s role in the oil spill (H1b), use thematic frames (H1c), or use unofficial sources (H1d).

Watson also found that journalists’ preferred professional roles—specifically their support for the Populist Mobilizer role—also affected journalists’ attitudes toward the oil industry.38 While this study also tests the indirect effect of these beliefs, journalists’ preferred professional roles might also have direct effects on journalists’ coverage of the BP oil spill.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 797

RQ1: Do journalists’ preferred professional roles predict how they covered the 2010 BP oil spill?

This study also examines the effect of community structure on journalists’ coverage of the BP oil spill. Previous studies have found that newspapers’ critical coverage of polluting industries was positively associated with structural pluralism and negatively associated with communities’ economic reliance on the oil industry.39

H2: Communities’ economic reliance on the oil industry will negatively predict the criticalness of journalists’ coverage of the BP oil spill; journalists in communities that are more reliant on the oil industry will be less likely to write negative stories (H2a), focus on BP’s role in the oil spill (H2b), use thematic frames (H2c), or use unofficial sources (H2d). H3: Communities’ degree of structural pluralism will positively predict the criticalness of journalists’ coverage of the BP oil spill; journalists in larger, more heterogeneous communities in which social power is likely to be more dispersed will be more likely to write negative stories (H3a), focus on BP’s role in the oil spill (H3b), use thematic frames (H3c), and use unofficial sources (H3d).

Previous studies have not examined how community structure’s influence enters the news-gathering process. This study tests Watson’s suggestion that these community-level variables might enter the news-gathering process by affecting journalists’ attitudes toward the oil industry, which in turn affect journalists’ coverage of the BP oil spill.40 If this hypothesis is correct, the community structure variables’ indirect effect on journalists’ coverage will be stronger than the direct effects.41

H4: The indirect effects of communities’ economic reliance on the oil industry and structural pluralism on journalists’ coverage of the BP oil spill, via journalists’ attitudes toward the oil industry, will be larger than the direct effects.

Method Survey The America’s News database was searched for Gulf Coast newspaper stories published between April 20, 2010, the day of the Deep Water Horizon explosion, and September 20, 2010, the day after BP sealed the leaking oil well. Articles that used the key phrases “BP,” “oil spill,” or “Deep Water Horizon,” and whose headlines and lead paragraphs suggested the story was about the BP oil spill, were downloaded. Authors’ bylines were copied into an Excel spreadsheet. Reporters’ mailing and e-mail addresses were copied from their stories and newspapers’ websites. A total of 683 unique bylines with valid e-mail addresses were recorded. Starting November 5, 2010, journalists received a pre-notification letter introduc- ing the study by mail. The survey was e-mailed to the journalists one week later. Two hundred twenty (32.3%) journalists completed the survey.42

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 798 Journalism & Mass Communication Quarterly 91(4)

Survey Measures Political ideology. Political ideology was measured using a single-item, seven-point scale: “How would you characterize your political ideology, from left (1) to right (7)?”43 This question was used because it was believed that due to the professional norm of objectivity, journalists might react negatively to more direct, multi-item measures of their personal political beliefs (M = 3.41, SD = 1.09).

Environmental ideology. Environmental ideology was measured using the four items from the New Environmental Paradigm that Dunlap and colleagues found had the highest correlations with the full twelve-item scale44: “Humans are severely abusing the environment,” “The balance of nature is strong enough to cope with the impacts of modern industrial nations” (reverse coded), “The so-called ‘environmental crisis’ has been greatly exaggerated” (reverse coded), and “Humans will eventually learn enough about how nature works to be able to control it” (reverse coded). Respondents indicated their level of agreement with each statement on a five-point scale anchored by “strongly disagree” (1) and “strongly agree” (5) (Cronbach’s α = .68, M = 3.78, SD = 0.57).

Professional roles. Journalists’ preferences for the different professional roles were measured using seven questions selected from Weaver and colleagues’ survey of American journalists based on the questions’ face validity.45 Respondents indicated their level of agreement with each statement on a five-point scale anchored by “strongly disagree” (1) and “strongly agree” (5). These questions measured four potential journalistic roles: the Adversarialist (two items, r = .98), the Populist Mobilizer (three items, α = .78), the Interpreter, and the Disseminator. Roles were not mutually exclusive; that is, journalists could embrace multiple professional roles (see items and their means in Table 1).

Attitudes toward oil drilling. This portion of the survey adapted thirteen questions from public opinion surveys about oil drilling,46 energy policy and government regulation,47 and industry responsibility48 (see Table 2 for question wording and descriptive statistics). Respondents indicated how strongly they agreed with each statement on a five- point Likert-type scale anchored by “strongly disagree” (1) and “strongly agree” (5) (α = .79, M = 2.68, SD = 0.56).

Demographics. In addition to age, gender, and income, journalists were also asked whether they held a journalism degree, how many years they had been a journalist, their tenure at their current newspaper, their primary job function (reporting, commentary, or editing), what beat they were assigned to, and whether they had any special training in covering environmental or energy issues.

Content Analysis Respondents to the survey had written 1,829 stories, 1,000 of which were randomly sampled and coded. The large range (1-82) in the number of stories individual journalists

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 799

Table 1. Journalists’ Preferred Professional Roles.

M SD Adversarialist Journalists should always be critical 2.29 1.10 of corporations Journalists should always be critical 2.34 1.12 of government Total 2.31 1.11 Populist Mobilizer Journalists should give people an 4.10 0.81 opportunity to express their views Journalists should motivate people 3.41 0.98 to get involved in the community Journalists should point to possible 3.83 0.87 solutions to community problems Total 3.78 0.75 Interpreter Journalists should provide analysis of 4.69 0.66 complex problems Disseminator Journalists should concentrate on 3.64 0.90 serving the largest possible audience N 164

Note. Items measured on a five-point Likert-type scale, strongly disagree (1) to strongly agree (5).

had written made it difficult to devise a meaningful summary of each journalist’s coverage. Thus, each story was treated as a unique case. Individual- and community-level data were matched with the stories based on bylines and the name of the community where a newspaper was located. A variable that grouped articles by the same author controlled for the violation of the statistical assumption that each case represents an independent observation. Because stories were randomly sampled, not all journalists who completed the survey were represented in the final data set. The final data set included 164 unique authors. The individual characteristics of the reporters in the final data set did not differ significantly from those characteristics of all journalists who responded to the survey. The first four paragraphs substantially about the BP oil spill in each story were coded, based on the assumption that these first paragraphs typically include a story’s lead and nutgraph, which state the story’s purpose and summarize its main points.49 Thus, this study did not include all elements of the stories, but rather stories’ primary focus and stories’ most prominent elements. The author conducted the content analysis, but a second trained coder analyzed a randomly selected sample of 10% (N = 100) of the stories in order to establish

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 800 Journalism & Mass Communication Quarterly 91(4)

Table 2. Journalists’ Attitudes toward the Oil Industry, Descriptive Statistics.

M SD The collapse of the Deep Water Horizon oil platform and 3.43 0.90 the oil spill in the Gulf of Mexico was a rare accident. U.S. energy policy should continue to encourage production 3.27 0.90 of more domestic oil supplies. The United States should encourage exploration for new 3.09 0.93 offshore oil fields. U.S. regulators should allow continued offshore drilling for oil 3.07 0.86 at current levels. Tragic as the Deep Water Horizon accident was, we cannot 2.96 0.94 let it get in the way of developing domestic oil supplies. Oil drilling companies are generally concerned with limiting 2.84 0.90 their environmental impact. The oil drilling industry generally complies with government 2.82 0.86 regulation. Oil drilling companies are committed to protecting the public. 2.60 0.86 The government does not do enough to regulate the oil 2.40 0.86 drilling industry (reverse-coded). Government regulation of the oil industry is adequate. 2.38 0.89 Strict environmental laws and regulations cost too many jobs 2.26 0.78 and hurt the economy. Government regulation of business is necessary to protect 1.99 0.77 the environment (reverse-coded). U.S. energy should shift attention away from fossil fuels to 1.96 0.80 sources of renewable energy (reverse-coded). Total attitudes toward oil industry 2.68 0.56 N 164

Note. Items measured on a five-point Likert-type scale, strongly disagree (1) to strongly agree (5). intercoder reliability. Reliability was measured based on simple agreement and Krippendorff’s alpha. Intercoder reliability exceeded the minimum .80 level of agreement for all variables.50

Evaluative tone. Positive stories are those stories that emphasize desirable outcomes: for example, that the environmental impacts of the oil spill were not as severe as originally feared they would be. Negative frames are those that emphasize negative outcomes: for example, setbacks in BP’s efforts to cap the well. Neutral frames either did not indicate whether outcomes were either positive or negative, or mentioned both simultaneously (simple agreement = 89%; α = .80).

Episodic/thematic frames. Using Iyengar’s typology,51 episodic frames are those that focus on a single event: for example, the latest report of where tar balls had washed ashore. Thematic frames examine broader trends and implications beyond an isolated

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 801

incident: for example, a story that explores the potential long-term negative economic impacts of additional government regulation of the oil industry (95%; α = .84).

Story subject. Each story was coded based on whether it focused primarily on BP’s role in the spill or the spill’s effect on BP (such as the effects of the spill on the company’s share prices); the spill’s impact on the rest of the oil industry; the government’s role in the oil spill, including government regulation and oversight of the oil industry, government hearings related to the oil spill, and so on; the environmental or health impacts of the spill; the economic impacts of the spill; and environmental or energy policy issues related to the spill. Only the first subject mentioned (i.e., a story’s primary subject) in each story was coded (86%; α = .82).

Sources. Lastly, the dichotomous presence/absence of “official” and “unofficial” sources was coded. Official sources included BP and other oil industry representatives, elected officials, and “other” government officials, such as employees of the Coast Guard. Unofficial sources included members of non-profits, environmental activists, and volunteers; independent scientists and engineers; and the employees of local busi- nesses, including fishermen and oil industry employees not commenting in an official capacity (91%; α = .82).

Community Structure Structural pluralism. Structural pluralism was measured based on the distribution of the population across different social categories using Blau’s Diversity Index  R 2  Diversity=−1 ∑ Pi , where Pi is the proportion of the population in a given category/  i=1  group.52 The index measures the probability, on a 0-to-1 scale, that two individuals drawn at random with replacement from the population will be from two different social groups. Census data were drawn from the 2009 American Community survey (ACS) five-year estimates based on the county where a newspaper was based.53 Race was divided into two groups, white and non-white; education was divided into five groups, less than high school, high school graduate, some college (including associate’s degree), college graduate, and advanced degree (master’s, JD, PhD, etc.); income was also divided into five groups, residents with a household median income less than $19,999, $20,000-$34,999, $35,000-$49,999, $50,000-$74,999, and more than $75,000; and industries were divided into three groups based on the percentage of the working population employed, the goods-producing (construction and manufacturing), services-producing (information services, transportation, retail trade, wholesale trade, etc.), and agricultural sectors. Population was also used as an indicator of a community’s degree of structural pluralism. To preserve the 0-to-1 scale of Blau’s index so that the final structural pluralism index is more easily interpretable than it would be if the measures were all standardized, each community’s population was measured as a ratio of the population of Houston, Texas, the largest community in the sample.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 802 Journalism & Mass Communication Quarterly 91(4)

While population is perhaps the best indicator of structurally pluralistic communities in the United States,54 it is also highly correlated with circulation, which might also affect the dependent measures. Thus, to make sure that the most pluralistic communities are not just the largest, but those that are also the most diverse racially, economically, and so on, the five individual structural pluralism indicators were added together and divided by five to give each pluralism measure equal weight in the final index (M = 0.38, SD = 0.06).

Economic reliance on the oil industry. Following Griffin and Dunwoody, economic reliance on the oil industry was measured as the percentage of the local workforce employed in “mining, quarrying, and oil and gas extraction” industries,55 according to the ACS (M = 0.01, SD = 0.01).This measure covers a broader swath of employment outside the oil industry. Due to the oil industry’s dominance in the Gulf Coast region, however, it is assumed that most individuals in this category were employed in the oil drilling industry. For the sake of brevity, this measure is referred to as the percentage of the workforce employed in the oil industry throughout this study.

Results The average journalist in the final sample was a white (83.5%, n = 137) male (54.3%, n = 89) reporter (76.8%, n = 126), 43 years old (M = 43.12, SD = 11.18), who leaned slightly left politically (M = 3.41, SD = 1.09). He held a bachelor’s degree in journalism (69.5%, n = 114), had been a reporter for 20 years, and had worked at his current newspaper for 11 years. He made between $40,000 and $50,000 a year. This profile is very similar to Weaver and colleagues’ national survey of journalists.56 These reporters had some concern for the state of the environment (M = 3.77, SD = 0.57) and little affinity for the oil industry following the BP oil spill (M = 2.68, SD = 0.56). The Adversarialist role was least popular among journalists (M = 2.31, SD = 1.12); the Interpreter role received the most support (M = 4.74, SD = 0.59). As shown in Table 3, journalists’ political ideologies were positively associated with their attitudes toward the oil industry following the BP oil spill; more conservative journalists had more positive attitudes toward the oil industry (β = .335, p < .001). Journalists’ pro-environmental beliefs were negatively associated with their attitudes toward the oil industry (β = −.336, p < .001). Journalists’ preferred professional roles were not significantly associated with their attitudes toward the oil industry. Journalists’ attitudes toward the oil industry were also positively associated with the percentage of the local workforce employed in the oil industry. Journalists who worked in communities that relied more heavily on the oil industry for employment had more positive attitudes toward the oil industry (β = .247, p < .001). In regard to journalists’ coverage, overall they were significantly more likely to focus on the government’s (n = 318) over BP’s (n = 115) responsibility for the oil spill, χ2(1, N = 1000) = 202.86, p < .001. The least frequent subject in journalists’ coverage was environmental or energy policy (N = 13). Journalists were also significantly more likely to use episodic (n = 941) rather than thematic frames (n = 59), χ2(1, N = 1000)

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 803

Table 3. Standardized Coefficients of Path Analysis Predicting Gulf Coast Newspaper Journalists’ Attitudes toward the Oil Industry and Their Coverage of the BP Oil Spill.

Content analysis variables Journalists’ attitudes Focus Reliance on toward the Story on BP’s Thematic unofficial oil industry tone responsibility frames sources Attitudes toward .126* .002 .114 .040 the oil industry Political ideology .335*** Environmental −.336*** beliefs Professional roles Interpreter −.038 −.072 .098* .234* −.140** Disseminator −.054 −.076 −.057 −.071 .054 Adversary −.058 −.063 −.055 .069 . 097 Populist .186 .126 −.184*** .081 .080 Mobilizer Community structure % of workforce .247*** .098 .091 −.112 .082 employed in oil industrya Structural .131 −.124* .087 .065 .081 pluralisma

Note. χ2(20, N = 987) = 22.556, p = .311; RMSEA = .011; CFI = .977; TLI = .922. RMSEA = root mean square error of approximation; CFI = comparative fit index; TLI = Tucker–Lewis index. a% of the workforce employed in the oil industry and structural pluralism were log-transformed to correct for non-normality. *p < .05. **p < .01. ***p < .001.

= 774.40, p < .001, and to rely on official (n = 774) rather than unofficial sources (n = 386), χ2(1, N = 1000) = 635.20, p < .001. Approximately half the journalists’ stories were negative (n = 499), and approximately 37% (n = 368) were neutral; just 13% (n =133) were positive.

Hypotheses and Research Questions Data analysis was conducted using MPlus (version 6), a software that can handle cat- egorical dependent variables (i.e., story tone, BP linkage, thematic frames, and unofficial characters) and can group non-independent observations—in this instance, more than one story written by the same author—into clusters. Creating these clusters controls for violation of the statistical assumption of independent observations.57 Before conducting the multivariate analysis, thirteen cases that were missing data for the survey measures were deleted, the most conservative approach to dealing with

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 804 Journalism & Mass Communication Quarterly 91(4) missing data that affects fewer than 5% of the cases.58 Thus, the multivariate analysis included 987 newspaper stories and 156 unique authors (i.e., clusters). While the model was a good fit,59 it explained a relatively small proportion of the variance in journalists’ focus on BP (R2 = .095), thematic frames (R2 = .092), the tone of their stories (R2 = .062), or their use of unofficial characters (R2 = .045). The model did explain a much larger proportion of the variance in journalists’ attitudes toward the oil industry (R2 = .451). The results of the path model are shown in Table 3. H1 predicted that journalists’ positive attitudes toward the oil industry would negatively predict whether they wrote negative stories about the oil spill (H1a) and whether their stories focused on BP’s role in the oil spill (H1b), used thematic frames (H1c), or used unofficial sources (H1d). As shown in Table 3, only H1a was supported: journalists with more positive attitudes toward the industry wrote more positive stories about the oil spill (β = .126, p < .05). RQ1 asked whether journalists’ preferences for the different professional roles predicted how they covered the 2010 BP oil spill. As shown in Table 3, journalists who embraced the Interpreter role were less likely to use unofficial sources (β = −.140, p < .01), were more likely to use thematic frames (β = .234, p < .05), and more likely to focus on BP’s role in the oil spill (β = .098, p < .05). Journalists who embraced the Populist Mobilizer role were significantly less likely to focus on BP’s role in the oil spill (β = −.184, p < .001). The other roles were not significantly associated with journalists’ coverage. H2 predicted that communities’ economic reliance on the oil industry will negatively predict whether journalists wrote negative stories about the oil spill (H2a) and whether their stories focused on BP’s role in the oil spill (H2b), used thematic frames (H2c), or used unofficial sources (H2d). As shown in Table 3, the percentage of the local workforce employed in the oil industry was not significantly associated with journalists’ coverage of the BP oil spill. Thus, H2 is not supported. H3 predicted that communities’ degree of structural pluralism would positively predict whether journalists wrote negative stories about the oil spill (H3a) and whether their stories focused on BP’s role in the oil spill (H3b), used thematic frames (H3c), or used unofficial sources (H3d). As shown in Table 3, only H3a was supported. Journalists in larger, more pluralistic communities did write significantly more negative stories about the oil spill (β = −124, p < .05). H4 predicted that the indirect effects of communities’ economic reliance on the oil industry and structural pluralism on journalists’ coverage of the BP oil spill, via journalists’ attitudes toward the oil industry, would exceed the direct effects. Only structural pluralism had a direct effect on the tone of journalists’ stories (β = −.124, p < .05). The indirect effect of structural pluralism on the tone of journalists’ stories, via journalists’ attitudes toward the oil industry, can be calculated by taking the product of the path from structural pluralism to journalists’ attitudes (β = .131) and the path from journalists’ attitudes to story tone (β = −.124), which equals −.016. The direct effect is larger. Thus, H4 is not supported.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 805

Conclusion According to Niven,

There are hundreds of articles published every year in U.S. newspapers on media bias, most alleging that newspapers favor liberals and Democrats over conservatives and Republicans . . . [but] despite the importance of this coverage of bias and this belief in bias, scholarly research on this topic has been hamstrung by limitations of method.60

This study responded to the need for methodological advancement in how bias is measured, matching individual-level survey data with a content analysis of the stories individual journalists wrote. As compared to aggregate-level data, this individual-level survey and content analysis data establish a more methodologically sound link between journalists’ ideology and their coverage than previous studies of bias. This study’s results, though, should be interpreted cautiously. It examined one non- randomly selected subset of newspaper journalists, in one region, reporting on a specific topic. Nonetheless, this study corroborates what others have found: There is minimal partisan bias in journalists’ coverage. More conservative, less environmen- tally concerned journalists had more positive attitudes toward the oil industry and in turn wrote slightly more positive stories about the spill. Journalists’ attitudes toward the oil industry, however, did not affect whether journalists focused on BP’s role in the disaster, their use of thematic frames, or their use of unofficial sources. The fact that ideology was a strong predictor of journalists’ attitudes toward the oil industry suggests that, personally, these journalists certainly are not ideologically neutral. But whether it is due to professional norms that guide their work or the watchful eye of an editor, journalists’ personal attitudes by and large were not reflected in their coverage of the oil spill. Despite these research findings, the debate over bias in mainstream news coverage is unlikely to subside. On one side of the debate, there are too many politicians, com- mentators, and average Americans who perceive that mainstream journalists’ coverage is biased. On the other hand, American journalists claim their work can be objective.61 The public is likely to continue to be skeptical of such claims, because as suggested above, journalists are not ideologically neutral. That does not relate directly to journalists’ claims that their coverage is objective—in theory, one can have strong attitudes that do not affect one’s work. But this study did find, however small, trace evidence that journalists’ personal ideology did have an effect on the tone of their coverage. Perhaps American journalists’ insistence that their work is objective simply adds fuel to the fire of the “liberal media” debate, particularly when there are bits of evidence that journalists are not drones who can completely separate their professional coverage from their personal beliefs. There is, though, consistent evidence that individual-level ideological bias does not play a significant role in shaping journalists’ reporting. There are, however, other factors that shape the production of media content, which are worthy of researchers’ attention. Beyond ideology, this study found that journalists’ coverage is also shaped by the professional and social contexts in which they

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 806 Journalism & Mass Communication Quarterly 91(4) work. Journalists who more strongly endorsed the Interpreter professional role used more thematic frames, delving more deeply into the underlying causes and consequences of the BP oil spill. However, these Interpreters were also less likely to use unofficial sources who are most likely to raise critical perspectives, reinforcing officials’ explanations over independent, critical perspectives. There were also some contradictions in stories of those journalists who embraced the Populist Mobilizer role; that is, they saw it as their role to give readers a voice in the media and to encourage them to be actively involved in community affairs. These journalists were less likely to focus on BP’s role in the crisis (perhaps because they viewed the government as more credible, and as a representative of the public’s voice), but they were not more likely to use unofficial sources (i.e., give the public a voice in their coverage). While most previous work on journalists’ professional roles has been descriptive in nature,62 this study suggests journalists’ beliefs about their role in society can also have predictive value. Future studies might also examine how well journalists’ work matches up with their self-conceived roles. (One might expect, for example, that Populist Mobilizers would give greater voice to unofficial sources in their coverage.) The method used here of matching individual-level survey and content analysis data would be useful to such endeavors. In regard to the community context in which journalists work, journalists in larger, more heterogeneous (i.e., structurally pluralistic) communities, in which social power is more likely to be dispersed, appeared to be freer to write negative stories involving a powerful corporate interest, such as BP. Previous studies of structural pluralism’s effect on news coverage have not offered an explanation of how such a structural bias enters the news-gathering process.63 This study did not find support for the suggestion that structural pluralism shapes journalists’ attitudes toward issues in the news, which in turn affect journalists’ coverage (i.e., that journalists’ attitudes mediate the relationship between structural pluralism and coverage). It is possible the lack of this individual-level path suggests that what biases exist at the community level are more institutional—whether at the level of media outlets, or the profession as a whole—than individual in nature. Future research should further examine how the influence of community structure enters the news-gathering process. This study did not address more institutionalized forms of bias, which would not manifest in comparisons of individual journalists’ coverage. However, despite a methodologically better match between individual journalists’ personal beliefs and their coverage, this study found results very similar to other studies that have concluded there is little evidence of individual-level biases in mainstream journalists’ coverage.64 Thus, perhaps there is little value in continuing to look for small traces of individual- level bias in journalists’ coverage (though audiences’ perceptions of bias deserve further study).65 One might make similar conclusions about researching the role of community structure or journalists’ preferred professional roles based on the size of the effects observed in this study. However, if journalists are part of a very complex professional, social, economic, and political system, as previously suggested, one would expect that any one factor would have a small impact on journalists’ coverage. Further studies of community structure are worthwhile because there are

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 807 still important theoretical advances to be made in understanding through what means community structure influences the news-gathering process. And due to the descriptive nature of previous studies of journalists’ preferred professional roles, there are likewise advancements to be made in that arena.

Acknowledgment The author thanks associate editor Dr. Julie Andsager and the anonymous reviewers for their assistance in improving this manuscript.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. Colby Hall, “Andrew Breitbart on BP Oil Spill: ‘This Is Media Scandal First and Foremost,’” Mediaite (blog), June 1, 2010, http://www.mediaite.com/online/andrew-breitbart-bp-oil-spill-is-first-and-foremost-a-media-scandal/ (accessed August 22, 2014). 2. Mark Jurkowitz, Tom Rosenstiel, and Amy Mitchell, “A Year in the News 2010,” Project for Excellence in Journalism, 2011, http://stateofthemedia.org/2011/mobile-survey/a-year- in-news-narrative/ (accessed April 18, 2013). 3. Dennis T. Lowry and Jon A. Shidler, “The Sound Bites, the Biters, and the Bitten: A Two- Campaign Test of the Anti-incumbent Bias Hypothesis in Network TV News,” Journalism & Mass Communication Quarterly 75 (4, 1998): 719-29; 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 & Mass Communication Quarterly 72 (1, 1995): 33-44; Paul Waldman and James Devitt, “Newspaper Photographs and the 1996 Presidential Election: The Question of Bias,” Journalism & Mass Communication Quarterly 75 (2, 1998): 302-11. 4. Dave D’Alessio and Mike Allen, “Media Bias in Presidential Elections: A Meta-Analysis,” Journal of Communication 50 (4, 2000): 133-56; Robert M. Entman, “Media Framing Biases and Political Power: Explaining Slant in News Coverage of Campaign 2008,” Journalism 11 (4, 2010): 389-408. 5. Entman, “Media Framing Biases and Political Power”; Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: A Two-Campaign Test”; Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: An Analysis of Network TV News Bias”; Waldman and Devitt, “Newspaper Photographs.” 6. Phillip J. Tichenor, George A. Donohue, and Clarice N. Olien, Community Conflict & the Press (Beverly Hills, CA: SAGE, 1980). 7. Entman, “Media Framing Biases and Political Power.” 8. D’Alessio and Allen, “Media Bias in Presidential Elections.” 9. D’Alessio and Allen, “Media Bias in Presidential Elections,” 131.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 808 Journalism & Mass Communication Quarterly 91(4)

10. Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: A Two-Campaign Test”; Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: An Analysis of Network TV News Bias.” 11. David H. Weaver, Randal A. Beam, Bonnie J. Brownlee, Paul S. Voakes, and G. Cleveland Wilhoit, The American Journalist in the 21st Century (Mahwah, NJ: Lawrence Erlbaum, 2007). 12. Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: A Two-Campaign Test”; Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: An Analysis of Network TV News Bias.” 13. David Niven, “Objective Evidence on Media Bias: Newspaper Coverage of Congressional Party Switchers,” Journalism & Mass Communication Quarterly 80 (2, 2003): 311-26. 14. Waldman and Dewitt, “Newspaper Photographs.” 15. Thomas E. Patterson and Wolfgang Donsbach, “News Decisions: Journalists as Partisan Actors,” Political Communication 13 (4, 1996): 455-68. 16. Patterson and Donsbach, “News Decisions.” 17. Brendan R. Watson, “Ideologies Drive Journalists’ Attitudes toward Oil Industry,” Newspaper Research Journal 33 (spring 2012): 6-22. 18. Randal A. Beam, David H. Weaver, and Bonnie J. Brownlee, “Changes in Professionalism of U.S. Journalists in the Turbulent Twenty-First Century,” Journalism & Mass Communication Quarterly 82 (summer 2009): 277-98. 19. Weaver et al., The American Journalist. 20. Beam, Weaver, and Brownlee, “Changes in Professionalism”; Weaver et al., The American Journalist. 21. Watson, “Ideologies Drive Journalists’ Attitudes.” 22. Tichenor, Donohue, and Olien, Community Conflict. 23. Robert J. Griffin and Sharon Dunwoody, “Impacts of Information Subsidies and Community Structure on Local Press Coverage of Environmental Contamination,” Journalism & Mass Communication Quarterly 72 (2, 1995): 271-84; Robert J. Griffin and Sharon Dunwoody, “Community Structure and Science Framing of News about Local Environmental Risks,” Science Communication 18 (June 1997): 362-84; Robert J. Griffin, Sharon Dunwoody, and Christine Gehrmann, “The Effects of Community Pluralism on Press Coverage of Health Risks from Local Environmental Contamination,” Risk Analysis 15 (4, 1995): 449-58. 24. Griffin, Dunwoody, and Gehrmann, “The Effects of Community Pluralism.” 25. Griffin and Dunwoody, “Impacts of Information Subsidies.” 26. Watson, “Ideologies Drive Journalists’ Attitudes.” 27. Watson, “Ideologies Drive Journalists’ Attitudes.” 28. D’Alessio and Allen, “Media Bias in Presidential Elections”; Entman, “Framing Bias”; Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: A Two-Campaign Test”; Lowry and Shidler, “The Sound Bites, the Biters, and the Bitten: An Analysis of Network TV News Bias”; Waldman and Dewitt, “Newspaper Photographs.” 29. Linda Jean Kensicki, “No Cure for What Ails Us: The Media-Constructed Disconnect between Societal Problems and Possible Solutions,” Journalism & Mass Communication Quarterly 81 (spring 2004): 53-73. 30. Shanto Iyengar, “Framing Responsibility for Political Issues: The Case of Poverty,” Political Behavior 12 (March 1990): 19-40; Shanto Iyengar, Is Anyone Responsible?: How Television Frames Political Issues (Chicago: University of Chicago Press, 1991); Shanto Iyengar, “Framing Responsibility for Political Issues,” Annals of the American Academy of Political and Social Science 546 (July 1996): 59-70.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 Watson 809

31. Philip Solomon Hart, “One or Many? The Influence of Episodic and Thematic Climate Change Frames on Policy Preferences and Individual Behavior Change,” Science Communication 33 (August 2010): 28-51. 32. Robert M. Entman and Andrew Rojecki, “Freezing Out the Public: Elite and Media Framing of the U.S. Anti-nuclear Movement,” Political Communication 10 (2, 1993): 155-73. 33. Conrad Smith, “News Sources and Power Elites in News Coverage of the Exxon Valdez Oil Spill,” Journalism Quarterly 70 (June 1993): 393-403. 34. Douglas Blanks Hindman, Robert Littlefield, Ann Preston, and Dennis Neumann, “Structural Pluralism, Ethnic Pluralism, and Community Newspapers,” Journalism & Mass Communication Quarterly 76 (June 1999): 250-63. 35. Griffin, Dunwoody, and Gehrmann, “The Effects of Community Pluralism.” 36. Griffin and Dunwoody, “Impacts of Information Subsidies.” 37. D’Alessio and Allen, “Media Bias in Presidential Elections.” 38. Watson, “Ideologies Drive Journalists’ Attitudes.” 39. Griffin, Dunwoody, and Gehrmann, “The Effects of Community Pluralism”; Griffin and Dunwoody, “Impacts of Information Subsidies.” 40. Watson, “Ideologies Drive Journalists’ Attitudes.” 41. Watson, “Ideologies Drive Journalists’ Attitudes.” 42. The survey was approved by the Institutional Review Board (IRB) at the University of North Carolina-Chapel Hill, where the author was enrolled when he originally collected data. The identities of journalists who responded to the survey were to be kept confidential, but the survey was not anonymous, which allowed for the matching of individual-level survey and content analysis data. 43. Patterson and Donsbach, “News Decisions.” 44. Riley E. Dunlap, Kent D. Van Liere, Angela G. Mertig, and Robert Emmet Jones, “New Trends in Measuring Environmental Attitudes: Measuring Endorsement of the New Ecological Paradigm: A Revised NEP Scale,” Journal of Social Issues 56 (fall 2000): 425-42. 45. Weaver et al., The American Journalist. 46. “Poll: Support for More Offshore Oil Drilling Plummets,” CBS News, May 11, 2010, http://www.cbsnews.com/news/poll-support-for-more-offshore-oil-drilling-plummets/ (accessed May 7, 2013). 47. Toby Bolson and Fay Lomax Cook, “Public Opinion on Energy Policy: 1974-2006,” Public Opinion Quarterly 72 (summer 2008): 364-88. 48. Barbara M. Miller and Janas Sinclair, “A Model of Public Response to Marketplace Advocacy,” Journalism & Mass Communication Quarterly 86 (autumn 2009): 613-29. 49. Jan Johnson Yopp, Katherine C. McAdams, and Ryan M. Thornburg, Reaching Audiences: A Guide to Media Writing, 5th ed. (Boston: Allyn & Bacon, 2009). 50. Andrew F. Hayes and Klaus Krippendorff, “Answering the Call for a Standard Reliability Measure for Coding Data,” Communication Methods and Measures 1 (1, 2007): 77-89. 51. Iyengar, “Framing Responsibility for Political Issues: The Case of Poverty”; Iyengar, Is Anyone Responsible?; Iyengar, “Framing Responsibility for Political Issues.” 52. Peter M. Blau, Parameters of Social Structure: Approaches to the Study of Social Structure (NY: Macmillan, 1975); Peter M. Blau, Inequality and Heterogeneity (NY: Free Press, 1977). 53. “2005-2009 American Community Survey 5-Year Estimates,” U.S. Census Bureau, 2010, http://factfinder.census.gov/servlet/DatasetMainPageServlet?_program=ACS&_ submenuId=datasets_2&_lang=en (accessed May 7, 2013).

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 810 Journalism & Mass Communication Quarterly 91(4)

54. David Pearce Demers, “Structural Pluralism, Intermedia Competition and the Growth of the Corporate Newspaper in the United States,” Journalism Monographs 145 (June 1994). 55. Griffin and Dunwoody, “Impacts of Information Subsidies”; “2005-2009 American Community Survey.” More direct measures of communities’ economic reliance on the oil industry, such as the percentage of the local GDP contributed by the oil industry, are not available from the Bureau of Economic analysis, in order to keep confidential data about individual non-public companies, especially in communities that may have only one company in that category. “Regional Data: GDP & Personal Income,” Bureau of Economic Analysis, 2010, http://www.bea.gov/iTable/iTable.cfm?reqid=70&step=1&isuri=1&ac rdn=2 (accessed May 8, 2013). 56. Weaver et al., The American Journalist. 57. Tihomir Asparouhov and Bengt Muthen, “Multilevel Modeling of Complex Survey Data” (paper presented at the Joint Statistical Meeting, Seattle, WA, August, 8, 2006). 58. Barbara G. Tabachnick and Linda S. Fidell, Using Multivariate Statistics, 5th ed. (Boston: Allyn & Bacon, 2007). 59. The proposed model was a good fit for the journalists’ data, χ2(14) = 22.556, p = .311, root mean square error of approximation [RMSEA] = .011, comparative fit index [CFI] = .977, Tucker–Lewis index [TLI] = .922. A non-significant chi-square indicates a good model fit, as does an RMSEA value of less than .05 and CFI and TLI values greater than .90. Michael W. Browne and Robert Cudeck, “Alternative Ways of Assessing Model Fit,” Sociological Methods & Research 21 (November 1992): 230-58. Li-tze Hu and Peter M. Bentler, “Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria versus New Alternatives,” Structural Equation Modeling: A Multidisciplinary Journal 6 (1, 1999): 1-55. 60. Niven, “Objective Evidence on Media Bias,” 321. 61. Alex Jones, Losing the News: The Future of News that Feeds Democracy (NY: Oxford University Press, 2009). 62. Beam, Weaver, and Brownlee, “Changes in Professionalism”; Weaver et al., The American Journalist. 63. Griffin, Dunwoody, and Gehrmann, “The Effects of Community Pluralism”; Griffin and Dunwoody, “Impacts of Information Subsidies”; Tichenor, Donohue, and Olien, Community Conflict. 64. D’Alessio and Allen, “Media Bias in Presidential Elections.” 65. Yariv Tsfati, Natalie Jomini Stroud, and Adi Chotiner, “Exposure to Ideological News and Perceived Public Opinion Climate: Testing the Media Effects Component of Spiral-of- Silence in a Fragmented Media Landscape,” International Journal of Press/Politics 19 (January 2014): 3-23.

Downloaded from jmq.sagepub.com at Baskent Universitesi on December 20, 2014 JMQXXX10.1177/1077699014550087Journalism & Mass Communication QuarterlyKim et al. 550087research-article2014

Communicating Science and Crises

Journalism & Mass Communication Quarterly 2014, Vol. 91(4) 811–829 What Corporations Say © 2014 AEJMC Reprints and permissions: Matters More than What sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699014550087 They Say They Do? A jmcq.sagepub.com Test of a Truth Claim and Transparency in Press Releases on Corporate Websites and Facebook Pages

Bokyung Kim1, Seoyeon Hong2, and Glen T. Cameron3

Abstract This study explicates two strategic ways to enhance organizational transparency in the eyes of its public during a crisis. By using 2 (a corporate website vs. a corporate Facebook) × 2 (presence vs. absence of statement asserting factuality of an official statement) × 2 (presence vs. absence of sentences disclosing more detailed information) mixed-subjects experiment (N = 133), this study showed that a statement by a corporation claiming the official announcement as true affects public assessment of organizational transparency and reputation, and that participants’ favorable reputation perceptions prompted by the truth claim were not determined by different online media.

Keywords transparency, Facebook, crisis communication, reputation

With the increasing popularity of crisis communication and the substantial influence of social media in communicating with publics,1 public relations research has begun to pay attention to investigating conditions that influence public responses to organizations’

1Rowan University, Glassboro, NJ, USA 2Webster University, St. Louis, MO, USA 3Missouri School of Journalism, Columbia, MO, USA Corresponding Author: Bokyung Kim, Department of Public Relations & Advertising, College of Communication & Creative Arts, Rowan University, 103 Bozorth Hall, Glassboro, NJ, USA. Email: [email protected]

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Downloaded from http://www.elearnica.ir 812 Journalism & Mass Communication Quarterly 91(4) strategic communication.2 In times of crisis, stakeholders form perceptions on organizational reputation, which is an important asset to an organization.3 Specifically, stakeholders evaluate the organization’s reputation based on information from three different sources: information directly generated by the organization, mediated information, and secondhand information from other publics.4 A press statement created by an organization and distributed online is a highly accessible source of crisis information that most publics search for first.5 Thus, organizations need to assess the potential effects of such organization-generated information. In estimating the effects of organization-generated information on reputation, scholars suggest that transparent public relations efforts offer an organization reputational benefit.6 Still, there is a research gap on stakeholder evaluations of those transparency efforts in social media. Specifically, one difficulty that has hindered previous research on transparency has been a lack of clear differentiation between the concept of transparency itself, public trust in an organization, and outcomes of organizations’ transparent communication efforts. Guided by extensive literature on trust and transparency,7 this study seeks to explicate the concept of transparency as compared with public trust of an organization. Transparency, as used in journalism studies,8 refers to a party’s openness or open communication to pursue his or her trustworthiness. Thus, this study describes the established operational definition of transparency as either the passive availability of information or the active disclosure of information.9 A relevant but unexplored area is how public relations practice defines the concept of transparency and how much transparency publics expect from organizations. Specifically, it is important to explore whether an organization can enhance its transparency by offering information about its crisis to publics. This study experimentally tests the effects of two factors on the public’s evaluation of transparency efforts and organizational reputation: (1) a claim presenting more detailed information in terms of disclosure (a disclosure claim) and (2) a claim asserting the organization-generated information as fact (a truth claim). To understand social media’s effect on corporate transparency, the current research also explores ways in which people assess information from different online media. The findings of this study provide substantial implications for communication scholars and practitioners seeking to understand better how an organization’s transparency efforts during a crisis may lead to more favorable perceptions of its reputation and to greater public trust.

Literature Review The Concept of Transparency The word transparent comes from the Latin preposition “trans,” which connotes movement, and “parent,” which means visible. Thus, transparency means “allowing everything to be visible.”10 Journalism scholars define transparency as the openness of information to the general public.11 For example, Craft and Heim wrote that globalization and the spread of democracy, promoted by the increasing popularity of new media,

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 813

led to a world that is now more independent, forcing people to be much more transparent in explaining their actions.12 In line with this research, Mitchell and Steele equated transparency with disclosure in three areas in blogging: (1) the principles you hold, (2) the procedures you follow, and (3) the person you are.13 Quoting Kovach and Rosenstiel,14 Allen also argued that the concept of transparency is practiced through the disclosure of newsgathering methods to increase the reliability of news.15 Specifically, the main argument is that transparency not only requires journalists to provide enough information to the public, but also requires journalists surrounded by suspicion to reveal procedures or motivations about news gathering.16 Despite the fact that transparency in journalism refers to open communication with the audience, scholars highlight the distinction between transparency and public trust.17 While transparent communication or transparent behavior is the best way to build public trust, it does not equal either trust or truth itself.18 Craft and Heim also regarded transparency as the antecedent of public trust, because as “greater accountability may be achieved, greater credibility may be enhanced.”19 Therefore, a journalist’s transparency is one way to earn public trust among investors, consumers, and regulators because it enhances the accountability of information. Specifically, the public’s evaluation of a transparency behavior, such as disclosing newsgathering processes or inviting readers to news meetings, increases accountability (i.e., accountability refers to explanations of why transparency efforts were reasonable and acceptable).20 This argument regarding the distinction between transparency and trust becomes clear in Plaisance.21 The author defines the concept of transparency as “openness in communication” that “serves a reasonable expectation of forthright exchange when parties have a legitimate stake in the possible outcomes or effects of the sending or receiving of the message.”22 In doing so, Plaisance regards transparency as merely one notion of truth-telling and points out that openness of information could fail to reduce deception and ensure public trust.23 More importantly, extensive trust literature provides powerful theoretical underpinning regarding the distinction between the two concepts: trustworthiness and transparency. In general, studies explicated one’s trust as based in cognitive and affective processes.24 McAllister defines interpersonal trust as “the extent to which a person is confident in and willing to act on the basis of, the words, actions, and decisions of another.”25 McAllister’s study and definition of interpersonal trust illustrate the distinction between cognition-based and affect-based trust.26 While affect-based trust is grounded in an individual’s perception of others’ behavioral motives and ongoing social interactions with others, cognition-based trust is established based on provided information (i.e., information about how a trustee did behave in the past and available background information about the trustee). Lewis and Weigert also assert that individuals choose which entities they trust using evidence that shows someone or something can be trusted.27 In terms of these arguments, transparency explicated as open communication may be a necessary condition for earning other’s trust. In other words, information disclosed by an entity pursuing transparency can be the cognitive content for estimating trustworthiness.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 814 Journalism & Mass Communication Quarterly 91(4)

In the same vein, Mayer and colleagues28 presented a definition of trust and “the Proposed Model of Trust” with factors of perceived trustworthiness: ability (i.e., skills or competences), benevolence (i.e., “the perception of a positive orientation of the trustee toward the trustor”), and integrity (“the consistency of the party’s past actions, credible communications about the trustee from other parties, belief that the trustee has a strong sense of justice, and the extent to which the party’s actions are congruent with his or her words”).29 In his model, available information from the trustee serves as an important factor to modify the trustee’s estimated ability, benevolence, and integrity, resulting in increased trust from the trustor. In this regard, both transparency and trustworthiness could be assessed using cognitive reasoning and given information. Furthermore, the trustee’s openness is an important antecedent of his or her trustworthiness.30 In this regard, the concept of transparency, which is engaging in disclosure, openness, or behavior to pursue open communication, is a functional means to establish one’s trustworthiness.

The Concept of Organizational Transparency in Public Relations Because transparency could increase public trust, the next concern is to apply transparency to the context of public relations and explicate transparent organizational communication delivered to publics. However, only a limited number of studies have considered the effects of transparency on public relations practices.31 More importantly, public relations studies have not fully accounted for whether transparency refers to transparent communication rooted in respectful exchange and relationships or whether it refers to strategic behavior to earn public trust. Public relations studies emphasize the importance of mutual understanding and ongoing, transparent communication for maintaining organization–public relationships.32 For example, Grunig articulated four principles of crisis communication: relationship, accountability, disclosure, and symmetrical communication. He pointed out that organizational openness to a public helps build public trust in organization–public relationships.33 Jahansoozi also stated that transparency and trustworthiness in an organization are important for building a relationship with a community by providing mutual support, collaboration, and cooperation between the community and organization.34 A conceptual and operational definition of transparency rooted in relationships can become clearer through reviewing Rawlins’s study.35 In his early study, Rawlins argued that stakeholders would trust an organization if they regard it as being vulner- able, competent, having goodwill, and honest. Stakeholders would build and trust an organization–public relationship, overcoming their sense of vulnerability and believ- ing that the organization would not abuse their trust.36 In the same vein, an organization should trust its stakeholders enough to risk being transparent.37 Here, both the organization and public must earn trust and reduce uncertainty between groups to execute transparency. For Rawlins, transparency requires trust; and trust requires an organization–public relationship.38

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 815

In other words, relational transparency illustrates two-way communication between an organization and its public. That is, mutual trust comes from strong and respect- based organization–public relationship. To be more transparent, an organization as well as its public should trust each other and be confident that neither would deceive one another. Here, transparency requires the trust-based relationship, while trust also requires transparency within the organization–public relationship framework. However, strategic transparency focuses more on one-way communication between the public and an organization. Strategic transparency lets an organization become an information provider and lets the public be the receiver of a message. In this definition, publics can have greater or lesser trust in the organization based upon evidence and information the organization opens to them. Therefore, strategic transparency refers to open communication behavior as a necessary condition for earning one’s trust.39 Rawlins also explained publics could cognitively evaluate an organization based on its three transparency dimensions: sharing substantial information with stakeholders, enhancing its accountability by offering balanced information, and gaining participation in a communication process.40 The first two dimensions indicate strategic transparency operated in one-way communication between the organization and its public. Some scholars similarly highlight strategic transparency, stating that public trust in organizations could be maintained by practices to increase the visibility of its actions.41 Specifically, Sweetser operationalized lacking transparency as hiding the truth during an organization’s campaign. She conducted an experiment, testing the effects of disclosure (i.e., the participants watched a viral video with the purpose of promotion, disclosing that the company created the video) and nondisclosure (i.e., the participants watched the viral video, but the company denied it produced the viral video) on a public’s perceptions on a given organization.42 In her findings, perceived openness in dialogue was a strategic means to build trust in organizations.43 Guided by the perspectives of strategic transparency,44 this study summarizes the definition of transparent communication as an organization’s strategic behavior to make as much information available as possible to help publics reason and to make organizations responsible and accountable. Journalism studies and the concept of cognitive-based trust inform these two assumptions: (1) For an emerging crisis, the information generated by an organization (e.g., an official announcement) is one of potential tactics to increase its transparency; and (2) the disclosed information of the organization can be a factor affecting public perceptions.45 This approach helps public relations scholars understand the ways in which publics react to content with various levels of transparency, especially in times of crisis.

The Outcomes of Transparency Claims during a Crisis If organizational transparency refers to its open communication or disclosure of information to publics, it is tied to corporate reputation.46 Executing transparency enhances the organization’s accountability where complaints and issues, along with its respon- sibilities, become clear. This in turn affects publics’ evaluations of the corporation.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 816 Journalism & Mass Communication Quarterly 91(4)

As an attempt to identify the consequences of transparency efforts, Rawlins initially constructed measures of stakeholder evaluation of organizational transparency in terms of three transparency reputation traits (integrity, respect for others, and openness) and four transparency efforts (participation, substantial information, accountability, and secrecy) by measuring transparency with four instruments.47 The most intriguing finding from his study was that transparent communication led to greater reputation in terms of “having integrity, respecting others, being open, and being ethical.”48 In addition, Rawlins confirmed that public relations practitioners could employ transparent communication through sharing substantial information (i.e., releasing accurate, timely, complete, reliable, comparable, understandable, and relevant statements) to directly contributing to an organization’s reputation.49 However, the study did not specify how organizational transparency done with a strategic communication approach affects reputation.50 Less is known about factors that contribute to greater or lesser transparency of organizations. In this regard, the current study revisits the two different operationaliza- tions of transparency in journalism studies: “passive availability of information” (i.e., the state in which documents, statistics, procedures, motives, and intentions are open and accessible to public view) and active “disclosure of it” (i.e., a process for bringing more detailed information into public view).51 Similarly, an organization can be transparent when giving information by displaying information clearly for public view (i.e., availability perspective) or by disclosing detailed information (disclosure). In other words, transparency here means to operate in such a way that it is easy for the public to see the actions of organizations. Concerning the passive availability perspective, a self-asserted truth claim can be one way to make information visible. In general, scholars emphasize positive discourse as a necessary means to change perceptions of the organization in crisis.52 Specifically, they emphasize the need to lessen reputational damage by altering not only attributions of the crisis, but also perceptions of what the organization says and does during crisis.53 To alter perceptions, an organization can generate an official press statement to convey concern and to communicate responding strategies. In this type of situation, claims made in a release appraise the opening information as true. Rawlins bolsters the assumption that stating the information visibly prevents perceptions of secrecy.54 Moreover, Good and colleagues stated that a trustee’s claims about how the trustee will behave are one of critical antecedents of the trustee’s trustworthiness.55 In other words, a claim to assert organization-generated information as factual may be a strategic way to contribute to an organization’s transparent communication. In turn, such transparent communication can affect reputation perceptions. Thus, the following hypotheses are proposed:

H1a: Participants will perceive an organization as more transparent when the claim asserting its official announcement as fact is present than when the claim is absent. H1b: Participants are more likely to have a favorable perception of an organization when the claim asserting its official announcement as fact is present than when the claim is absent.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 817

However, the active disclosure perspective describes the efforts to present more detailed explanations.56 Consequently, the act of providing more explanation is perceived as greater transparency and leads to a more favorable reputation.57 Quoting the Public Relations Society of America code of ethics provision on disclosure of information, Plaisance also supports this assertion, stating that public trust can be established “by revealing all information needed for responsible decision-making.”58 Therefore, H2a and H2b, illustrating the impact of presenting more detailed information on the outcomes of transparent communication, are proposed:

H2a: Participants are more likely to evaluate an organization presenting detailed information on its official announcement as more transparent than when the information is absent. H2b: Participants are more likely to have favorable perception of an organization presenting detailed information on its official announcement than when the information is absent.

However, either an official announcement claiming factuality or an announcement providing more detailed information may not enhance an organization’s perceived transparency.59 Allen differentiated truth, defined as “the transparent nature of dialogue,” from transparency efforts.60 O’Neill also indicated that a message sender’s transparency and openness can produce a flood of information that may cause confu- sion.61 In other words, claiming truthfulness or providing more information in an official announcement can merely increase the flow of information without enhancing meaning.62 Based on this argument, the first research question regarding an interaction effect of the two claims on our dependent variables is as follows:

RQ1: Does a statement claiming facts of its official announcement with a disclosure of more detailed information have an effect on the participants’ (1) perceptions of an organization’s reputation and (2) evaluations of its transparency efforts?

Social Media Effect on Transparency Perception Understanding when people call for greater or lesser transparency in different media is another concern related to transparent communication. Coombs pointed out that crisis information originating from an organization is one of the major channels for publics to gather information during a crisis.63 It follows that an organization cannot guarantee the crisis news coverage including its official statement. It also cannot control crisis information created by private individuals who share their views with other publics. However, the organization can control its official announcement and the stories appearing on its own website and Facebook. Stephens and Malone also argued that organizational websites serve as highly accessible channels for organizations that can reach out to a variety of stakeholders in crises.64 Such online channels play an important role in disseminating the organization’s

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 818 Journalism & Mass Communication Quarterly 91(4) information. Sweetser also asserted that organizations have to deal with noncustomers and those who could access online information provided by the organization.65 Walton also stated “Web 2.0 forced transparency upon organizations,” claiming that publics expect more detailed information in social media.66 This raises the question of whether publics evaluate an organization’s transparency claim differently if it is displayed on the corporate website or on corporate Facebook. Consequently, guided by empirical research concerning the channel effect, the second question is generated.

RQ2: Does online channel have an effect on the participants’ (1) perceptions of an organization’s reputation and (2) evaluations of its transparency efforts?

Method Research Design This study used a 2 (online channel: a corporate website vs. a corporate Facebook) × 2 (statement claiming an official statement as true facts: presence vs. absence) × 2 (claims to present more detailed information: presence vs. absence) × 2 (message order) posttest-only mixed-subjects experimental design. Online channel was a within- subject factor. Thus, each participant read two news releases that appeared on both a corporate website and Facebook. Claims presenting more detailed information and asserting the official statement as true, and the order of the within-subject factor, were between-subject variables. Participants were assigned to one of four different conditions. In the first condition, a disclosure claim was presented in response to the stakeholders’ request (i.e., presenting more information condition) asserting that the claim is a fact (i.e., a self-asserted truth claim). In the second condition, such claims were presented without the truth claim. In the third condition, any detailed information regarding the stakeholders’ request was absent, while presenting the truth claim. In the last condition, both truth- claiming statements and detailed information were not shown. Thus, participants in each group read different versions of an organization’s official statement appearing on the corporate website and corporate Facebook.

Participants Researchers recruited 133 students from undergraduate courses in different departments at a midwestern university. College students were considered appropriate participants in the current study because the online channels used as stimuli are popular among young adults. A larger percentage of participants were females (69.9%, n = 93) with an average age of 21.33. The majority of participants were Caucasian (70.7%, n = 94), while 14.3% (n = 19) were Asian, 6.8% (n = 9) were African American, 3.8% (n = 5) were Hispanics, 1.5% were Native Americans (n = 2), and 1.5% identified themselves as Others (n = 2).

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 819

Stimuli The researchers selected the Taco Bell crisis (i.e., lawsuits over the real ingredients of Taco Bell’s products) for stimulus scenarios. Taco Bell was accused of false advertising about its “meat” mixture containing binders and extenders, which did not meet the standards of the U.S. Department of Agriculture (USDA) for product labeling of meat. The accuser did not suffer from medical complications, but he wanted Taco Bell to be honest in its advertising. In response to this lawsuit, Taco Bell provided its “gratitude statement” on its official website and revealed the ingredients in its products. We used this official announcement as our stimulus. Concerning the within-subject factor, the same contents in two different versions were provided for the online channel effect: news releases on Facebook and the corporate website. In terms of the between-subject variable, two versions of news releases were created with similar wording, but with different claims of detailed procedures. The dimension of both claims was also manipulated by either adding one or two sentences regarding each factor or excluding such sentences. For example, “So, here is the truth about our beef, 88% beef and 12% seasonings” was used as a truth claim, while noting “we are sharing our actual recipe” in the disclosing information condition. To match the same design and word count, we included a short history of Taco Bell to replace manipulated sentences. Dimension of the online channel was manipulated by differentiating the actual platform of Taco Bell’s official announcement presented on Facebook and its official website, and everything was identical, except the manipulated two transparency claims as the independent variable.

Measures Prior reputation. Prior to viewing an organization’s news releases, participants were asked to rate their perceived baseline reputation of the stimuli company (e.g., Taco Bell) and other four companies (e.g., McDonalds, Burger King, Wal-Mart, and Chipo- tle). The participants’ prior reputation was measured with one item developed by Coombs and Holladay.67 The item is, “Overall, my impression of ‘the company X’ is . . .” on a scale ranging from very unfavorable to very favorable.

Organizational reputation. The organizational reputation was measured using the 5-item Organizational Reputation Scale developed by Coombs and Holladay68 (α = .94).

Perceptions of transparency efforts. The 17-item Communication Efforts Scale by Raw- lins measured the participants’ evaluation of an organization’s transparency efforts.69 Among the original four constructs of transparency scales, this study excluded the participation components for two reasons: (1) The participants’ mere exposure to particular information does not reflect the participatory nature of organizational transparency; and 2) according to the previous study regressing overall transparency evaluation on its four components, if participation is excluded in a regression model, substantial

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 820 Journalism & Mass Communication Quarterly 91(4) information, accountability, and secrecy together significantly predict overall organizational transparency.70 Thus, the scale comprises the remaining three constructs: (1) Substantial information, (2) Accountability, and (3) Secrecy (α = .96).

Results Manipulation Check All manipulations worked as intended. To measure presence of the self-asserted truth claim, the participants were asked to indicate their level of agreement on a 7-point Likert-type scale (1 = strongly disagree; 7 = strongly agree) with whether the official statement they just read included true facts claimed by Taco Bell. For measuring the condition of presenting more information, the participants were also asked to rate their level of agreement on a 7-point Likert-type scale (1 = strongly disagree; 7 = strongly agree) with whether the announcement they viewed included the company’s disclosure regarding its recipe. The manipulation check was tested using independent samples t tests. The results indicate that participants in the condition of truth claims significantly agreed more that the announcement they read included sentences insisting the fact is truth (M = 5.58, SD = 1.28) compared with participants in the condition absent of such claims (M = 4.873, SD = 1.53), t(131) = −3.50, p < .01. Similarly, participants in the condition of presenting more information significantly agreed more that the statement they read included the company’s disclosure regarding its detailed recipe (M = 5.08, SD = 1.68) than participants in the condition absent of such detailed information (M = 3.51, SD = 1.84), t(131) = −5.14, p < .001.

Hypothesis Testing H1a and H1b predicted that the participants would report greater perceived transparency of an organization (H1a) and more favorable perceptions of an organization’s reputation (H1b) in the condition of claiming truthfulness of its official announcement. The effect of the self-asserted claim to true fact on organizational transparency and reputation perception was tested using multivariate analysis of covariance (MANCOVA) with the organization’s prior reputation as a covariate. The present experimental design included three dimensions of independent variables: two between- subjects factors (i.e., a truth claim and a disclosure claim) with two levels, one within- subjects factor (i.e., online channel) with two levels, and the participants’ prior reputation of stimuli organization (e.g., Taco Bell) as a control variable. The Pearson correlation indicated that a moderate, positive, and significant relationship exists between prior reputation of Taco Bell and postorganizational reputation after reading its official statement, r(131) = .43, p < .001; and a strong, positive, and significant relationship exists between the company’s prior reputation and its perceived efforts to pursue transparency, r(131) = .86, p < .001. Based on these procedures, the

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 821

organization’s prior reputation served as a control variable because this variable is assumed to influence the two dependent variables and thus needed to be adjusted to obtain a more precise analysis. The findings from the MANCOVA revealed that the effect of the truth claim was significant on the two dependent variables, Wilks’s Λ= .852, F(2, 129) = 11.24, p 2 <.001, ηp = .148. The univariate analyses also showed that the main effect of viewing the self-asserted claim to true fact was significant for the participants’ perceived orga- 2 nizational transparency, F(1, 130) = 22.52, p < .001, ηp = .148, and their reputation 2 perceptions on the given company, F(1, 122) = 14.24, p < .001, ηp = .099. These results suggest that viewing sentences claiming factuality of the official announcement were more likely to make participants evaluate Taco Bell as transparent (M = 4.60) and perceive its reputation as favorable (M = 4.41) than those who read the announcement

without truth claims (Morganizational transparency = 3.88 and Mreputation = 3.74). Therefore, H1a and H1b were supported. H2a and H2b predicted that the participants’ perceptions of organizational transparency and its reputation would be affected by the conditions with more detailed information. The effect of the disclosure claims on the two dependent variables was tested using MANCOVA with the organization’s prior reputation as a covariate. The results of MANCOVA indicated that the effect of disclosing detailed information did not have a statistically significant effect on the participant’s perceptions of the given 2 company, Wilks’s Λ= .994, F(2, 129) = .42, p = .66, ηp = .006. According to subsequent univariate analyses, Taco Bell’s official announcement with more information presented (M = 4.14, SE = .11) and its official announcement without disclosure (M = 4.02, SE = .11) were not statistically different in terms of participants’ perceptions of 2 the organization’s transparency, F(1, 130) = .628, p = .43, ηp = .005. Similarly, the official announcement with the detailed information present (M = 4.28, SE = .15) and the announcement without the disclosure (M = 4.20, SE = .15) were not statistically different in terms of participants’ perceptions of organizational reputation, F(1, 130) = 2 .193, p = .66, ηp = .001. Thus, H2a and H2b were not supported. RQ1 asked whether a statement claiming factuality of the official announcement interacted with a claim offering more details on the two dependent variables. The results of MANCOVA showed that the effect of the self-asserted claim of true fact did not significantly interact with the effect of presenting more detailed information, 2 Wilks’s Λ = .992, F(2, 127) = .502, p = .61, ηp = .008. However, the combined dependent variables of perceived transparency of an organization, F(1, 128) = 22.10, p < η2 2 .001, p = .148, and its reputation, F(1, 128) = 14.02, p < .001, ηp = .099, were still significantly affected by the truth claim, Wilks’s Λ = .852, F(2, 127) = 11.03, p < .001, while accounting for 15% of the variability in the dependent variables after controlling for the participants’ baseline reputation for Taco Bell. The results of the subsequent univariate analyses also revealed that the participants’ transparency perceptions of an 2 organization, F(1, 128) = .43, p = .84, ηp = .003, and its reputation, F(1, 128) = .51, p 2 = .48, ηp = .003, were not significantly affected by the interaction effect between the two claims.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 822 Journalism & Mass Communication Quarterly 91(4)

Figure 1. Interaction effect between truth claims and online channel on organizational reputation.

RQ2 asked whether online channels would influence the participants’ perception of an organization’s reputation and evaluations of its transparency efforts. A repeated measures of MANCOVA with Taco Bell’s baseline reputation as a covariate showed that the combined dependent variables of perceived transparency of an organization and its reputation were not significantly affected by the conditions of online channels, 2 Wilks’s Λ = .978, F(2, 127) = 1.44, p = .24, ηp = .022; but were significantly affected 2 by the self-asserted truth claim, Wilks’s Λ = .978, F(2, 127) = 11.03, p < .001, ηp = .148. In addition, there was a statistically significant interaction effect between the conditions of the truth claim and online platforms on the combined dependent vari- 2 ables, Wilks’s Λ = .943, F(2, 127) = 3.86, p < .05, ηp = .057.

Additional Analysis The interactions of the two independent variables were analyzed in addition to main hypotheses testing. The following univariate analyses showed that the interaction between conditions of claims stressing true fact and of online channels was found to be significant only for the participants’ evaluations of organizational reputation, F(1, 2 128) = 6.74, p < .05, ηp = .050; but not significant for perceived transparency of the 2 company, F(1, 128) = .02, p = .96, ηp = .001. Specifically, in the condition of the absent truth claim, participants reported more favorable corporate reputation when viewing the company’s announcement posted on a corporate website (M = 4.03, SE =

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 823

.14) compared with when reading the same statement on its Facebook page (M = 3.73, SE = .15). Conversely, in the condition of the present truth claim, participants’ reputation perceptions toward the given company would not differ much when viewing the statement on Facebook (M = 4.62, SE = .15) or when reading it on the corporate website (M = 4.59, SE = .14; see Figure 1).

Discussion and Conclusion Audiences tend to see a company as more transparent and as having a higher reputation when they read a statement claimed to be true by an organization versus a statement without truth claims. This argument is crucial in that strategic transparency is open communication that results in two claim types, as operationalized in the context of the organization-constructed message during a crisis. An organization’s production of claims to describe its behavior could be seen as one potential way to enhance organizational transparency. This practice could lead to public trust in the organization, as supported by Good.71 Good uses a metaphor to posit that limits of publics make them rely on provided claims and information:

If I habitually get my groceries at one store because it claims to be the cheapest and initially it seemed to be so, and the prices increase from week to week, I can take this as evidence that inflation is at a certain level, and continue to shop there.72

In this example, Good emphasizes that our cognitive processing limits intake of information, forcing us to make decisions based on given information.73 In other words, people prefer to have a manageable amount of information, while they acknowledge the flaw of such reduction of information.74 As this study manipulated press releases to differ only by the online medium and the claims, we could assume that an organization’s statement asserting factuality could produce an information limitation. This might have led the participants to focus on the claim stressing truth, and base their future actions and perceptions on the available claim. This logic connects the truth or factuality claims to readers’ cognitive-based trust.75 Another major finding is that participants’ favorable reputation perceptions after reading the truth claim were not determined by medium, whereas the same was not true for the statement without a truth claim. Specifically, when the same message content appeared without a truth claim, the corporate website resulted in more favorable reputation perceptions toward the company than a corporate Facebook page. However, when the company highlighted the factuality of its announcement, the corporate website and Facebook did not produce different results for participants’ corporate reputation perceptions. In other words, when organizations asserted a corporate announcement as true in a claim, the different reputation perceptions prompted by the different online media were minimized. Regardless of various online channels, the participants had more favorable reputation perceptions when they had the truth claim compared with an absent truth claim. Thus, the truth claim appears to have a powerful impact on the participants’ reputation perceptions toward organizations, even when it is on Facebook.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 824 Journalism & Mass Communication Quarterly 91(4)

Scholars insist that various social media channels be more conversational and relationship-oriented, because they are venues where people can interact with others.76 Thus, they may be less reluctant to accept truth claims appearing on Facebook where the claim feels like assurance from one interpersonal participant to another.77 Nevertheless findings show that presenting the truth claim created favorable reputation perceptions toward the company regardless of the online medium. We may conclude that any online platforms may create a context where assurances of the truth-telling are considered appropriate. Furthermore, this study offers practical implications regarding the value of the truth claim. It is worth mentioning that our participants reported a relatively lower prior reputation score for the target company (M = 4.08), compared with four other corporations (M = 4.42), t(132) = −2.49, p < .05. This allowed more latitude to interpret the findings. The participants could not have formed trust relationships with the stimulus company, Taco Bell. The truth claim may help publics without pre-existing trust relationships estimate the company as transparent and favorable. Surely, public relations practitioners cannot eliminate stakeholders’ doubts solely by a claim or statement. However, public relations practitioners can construct and release crisis information (e.g., an official statement) to be more transparent by stressing the information as true fact, in the way that our data suggest. The organization’s use of its online presence with a goal to reach out to a mass audience within the online platform can especially help transparency in this way. Although our manipulation check for the disclosure statements was successful, the statements were not sufficient for increasing transparency perceptions toward the organization. People did not perceive the official statement with Taco Bell’s detailed recipe of its seasoned beef as being more transparent than the official statement with the company’s history instead of the recipe details. One possible explanation is that these two different conditions were not conceptually different in terms of perceived transparency efforts of the organization and its reputation. If the perceivers do not have mutual respect and existing relationships with the reputation holder, a relatively greater amount of facts in the disclosure condition would either merely increase the amount of information or create an obfuscating blizzard of content and not increased transparency.78 However, this finding also presents positive implications to practitioners. Practically speaking, presenting an organization’s history could be another option for public relations practitioners in designing official releases. Similar to other experimental studies, this study was limited in that we relied on a convenience sample of college students. However, the organization used in the study targets young consumers, including college students. Basil and colleagues also ascer- tained that the use of student samples does not violate the validity of logical conclusion, especially when a researcher aims to test multivariate relationships among variables of interest.79 Another limitation of this study is related to the dissimilar nature of disclosure claims. Whittington and Yakis-Douglas provide an insight: two types of disclosure exist, either a voluntary disclosure or more mandatory disclosure.80 Specifically, they showed that if people thought the information disclosure was mandatory and

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 825

inevitable for an organization, the effect of the disclosure would not be critical. In addition, scholars cumulatively have demonstrated the effectiveness of the proactive self-disclosure strategy—called “stealing thunder”—on publics’ behaviors and perceptions.81 In terms of the sentence in the condition presenting detailed information (e.g., “in response to a consumer request, we are sharing our actual recipe”), the participants might regard this condition as more mandatory disclosure than voluntary disclosure. The perceivers of a reputation might see an official statement providing an organization’s history as having the same significance as the statement presenting detailed information about the crisis as well. Future research might include the different levels of disclosure (active stealing thunder or nonthunder), as well as the different amount of information regarding organizations’ ability, benevolence, and integrity, which in turn affect public trust in organizations.82 Furthermore, it would be fruitful to know the effects of the truth claim in traditional media, such as newspapers, on public perceptions and behaviors. Still, this study breaks new ground by experimentally testing the effects of the two types of organizational claims on two important crisis outcome variables: perceived organizational transparency and reputation. Transparency has been understudied as a major factor in the acceptance of public relations messages. To build upon transparent communication research, it will become increasingly important to explore the inter- play between truth claims, open communication, and transparency of an organization.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes 1. Kerl K. Stephens and Patty C. Malone, “If the Organizations Won’t Give Us Information: The Use of Multiple New Media for Crisis Technical Translation and Dialogue,” Journal of Public Relations Research 21 (April 2009): 229-39; Donald K. Wright and Michelle D. Hinson, “How Blogs and Social Media Are Changing Public Relations and the Way It Is Practiced,” Public Relations Journal 2 (spring 2008): 1-21. 2. Kaye D. Sweetser, “A Losing Strategy: The Impact of Nondisclosure in Social Media on Relationships,” Journal of Public Relations Research 22 (July 2010): 288-312. 3. Timothy W. Coombs, “Protecting Organization Reputations during a Crisis: The Development and Application of Situational Crisis Communication Theory,” Corporate Reputation Review 10 (September 2007): 163-76. 4. Timothy W. Coombs and Sherry J. Holladay, “The Negative Communication Dynamic: Exploring the Impact of Stakeholder Affect on Behavioral Intentions,” Journal of Communication Management 11 (November 2007): 300-12.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 826 Journalism & Mass Communication Quarterly 91(4)

5. Coombs, “Protecting Organization Reputations during a Crisis”; Kerl K. Stephens and Patty C. Malone, “New Media for Crisis Communication: Opportunities for Technical Translation, Dialogue, and Stakeholder Responses,” in Handbook of Crisis Communication, ed. Timothy W. Coombs and Sherry J. Holladay (Malden, MA: Blackwell, 2010), 381-95. 6. Brad L. Rawlins, “Give the Emperor a Mirror: Toward Developing a Stakeholder Measurement of Organizational Transparency,” Journal of Public Relations Research 21 (January 2009): 71-99; Sweetser, “A Losing Strategy.” 7. David S. Allen, “The Trouble with Transparency: The Challenge of Doing Journalism Ethics in a Surveillance Society,” Journalism Studies 9 (June 2008): 323-40; John K. Butler, “Toward Understanding and Measuring Conditions of Trust: Evolution of a Conditions of Trust Inventory,” Journal of Management 17 (September 1991): 643-63; Stephanie Craft and Kyle Heim, “Transparency in Journalism: Meanings, Merits, and Risks,” in The Handbook of Mass Media Ethics, ed. Lee Wilkins and Clifford G. Christians (NY: Routledge, 2009), 217-28; George F. Farris, Eldon E. Senner, and D. Anthony Butterfield, “Trust, Culture, and Organizational Behavior,” Industrial Relations: A Journal of Economy and Society 12 (May 1973): 144-57; John Gabarro, “The Development of Trust, Influence, and Expectations,” in Interpersonal Behavior: Communication and Understanding in Relationships, ed. Anthony G. Athos and John J. Gabarro (Englewood Cliffs, NJ: Prentice Hall, 1978), 290-303; Kerry M. Hart, Randall Capps, Joseph P. Cangemi, and Larry M. Caillouet, “Exploring Organizational Trust and Its Multiple Dimensions: A Case Study of General Motors,” Organization Development Journal 4 (summer 1986): 31-39; Julia Jahansoozi, “Organization-Stakeholder Relationships: Exploring Trust and Transparency,” Journal of Management Development 25 (October 2006): 942-55; Carl I. Hovland, Irving L. Janis, and Herold H. Kelley, Communication and Persuasion (New Haven, CT: Yale University Press, 1953); Daniel J. McAllister, “Affect- and Cognition-Based Trust as Foundations for Interpersonal Cooperation in Organizations,” Academy of Management Journal 38 (February 1995): 24-59; Patrick L. Plaisance, “Transparency: An Assessment of the Kantian Roots of a Key Element in Media Ethics Practice,” Journal of Mass Media Ethics 22 (December 2007): 187-207; Brad L. Rawlins, “Measuring the Relationship between Organizational Transparency and Employee Trust,” Public Relations Journal 2 (spring 2008): 1-21; Rawlins, “Give the Emperor a Mirror.” 8. Allen, “The Trouble with Transparency”; Craft and Heim, “Transparency in Journalism”; Plaisance, “Transparency.” 9. Allen, “The Trouble with Transparency”; Craft and Heim, “Transparency in Journalism”; Bill Mitchell and Bob Steele, “Earn Your Own Trust, Roll Your Own Ethics, Transparency and Beyond,” (January, 2005), Paper presented at the Blogging, Journalism and Credibility Conference, Harvard University, Cambridge, MA, http://cyber.law.harvard.edu/webcred (accessed January 17, 2011). 10. Plaisance, “Transparency,” 205. 11. Allen, “The Trouble with Transparency”; Craft and Heim, “Transparency in Journalism”; Mitchell and Steele, “Earn Your Own Trust”; Plaisance, “Transparency.” 12. Craft and Heim, “Transparency in Journalism.” 13. Mitchell and Steele, “Earn Your Own Trust.” 14. Bill Kovach and Tom Rosenstiel, The Elements of Journalism: What News People Should Know and the Public Should Expect (NY: Crown, 2001). 15. Allen, “The Trouble with Transparency.” 16. Allen, “The Trouble with Transparency.”

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 827

17. Allen, “The Trouble with Transparency”; Craft and Heim, “Transparency in Journalism”; Plaisance, “Transparency.” 18. Allen, “The Trouble with Transparency.” 19. Craft and Heim, “Transparency in Journalism,” 222. 20. Craft and Heim, “Transparency in Journalism.” 21. Plaisance, “Transparency.” 22. Plaisance, “Transparency,” 188. 23. Plaisance, “Transparency.” 24. J. David Lewis and Andrew Weigert, “Trust as a Social Reality,” Social Forces 63 (June 1985): 967-85; McAllister, “Affect- and Cognition-Based Trust.” 25. McAllister, “Affect- and Cognition-Based Trust,” 25. 26. McAllister, “Affect- and Cognition-Based Trust.” 27. Lewis and Weigert, “Trust as a Social Reality.” 28. Roger C. Mayer, James H. Davis, and F. David Schoorman, “An Integrative Model of Organizational Trust,” Academy of Management Review 20 (June 1995): 709-34. 29. Mayer, Davis, and Schoorman, “An Integrative Model of Organizational Trust,” 719. 30. Butler, “Toward Understanding and Measuring Conditions of Trust”; Farris, Senner, and Butterfield, “Trust, Culture, and Organizational Behavior”; Gabarro, “The Development of Trust, Influence, and Expectations”; Hart et al., “Exploring Organizational Trust and Its Multiple Dimensions”; Hovland, Janis, and Kelley, Communication and Persuasion. 31. Rawlins, “Give the Emperor a Mirror”; Rawlins, “Measuring the Relationship”; Sweetser, “A Losing Strategy.” 32. James E. Grunig and Yi-Hui Huang, “From Organization Effectiveness to Relationship Indicators: Antecedents of Relationships, Public Relations Strategies, and Relationship Outcomes,” in Public Relations as Relationship Management, ed. John A. Ledingham and Stephen D. Bruning (Mahwah, NJ: Lawrence Erlbaum, 2001), 23-53; John A. Ledingham and Stephen D. Bruning, “Organization and Key Public Relationships: Testing the Influence of the Relationship Dimensions in a Business-to-Business Context,” in Public Relations as Relationship Management: A Relational Approach to the Study and Practice of Public Relations, ed. John A. Ledingham and Stephen D. Bruning (Mahwah, NJ: Lawrence Erlbaum, 2000), 159-73. 33. James E. Grunig, “Two-Way Symmetrical Public Relations: Past, Present, and Future,” in Handbook of Public Relations, ed. Robert. L. Heath (Thousand Oaks, CA: SAGE, 2001), 11-30. 34. Jahansoozi, “Organization-Stakeholder Relationships.” 35. Rawlins, “Measuring the Relationship.” 36. Rawlins, “Measuring the Relationship.” 37. Rawlins, “Measuring the Relationship.” 38. Rawlins, “Measuring the Relationship.” 39. Rawlins, “Measuring the Relationship.” 40. Rawlins, “Measuring the Relationship.” 41. Sweetser, “A Losing Strategy.” 42. Sweetser, “A Losing Strategy.” 43. Sweetser, “A Losing Strategy.” 44. Sweetser, “A Losing Strategy”; Rawlins, “Measuring the Relationship”; Rawlins, “Give the Emperor a Mirror.” 45. Coombs and Holladay, “The Negative Communication Dynamic.”

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 828 Journalism & Mass Communication Quarterly 91(4)

46. Rawlins, “Give the Emperor a Mirror.” 47. Public Relations Society of America, “Public Relations Society of America Member Code of Ethics,” http://prsa.org/aboutUs/ethics/preamble_en.html (accessed January 28, 2004); Rawlins, “Give the Emperor a Mirror.” 48. Rawlins, “Give the Emperor a Mirror,” 94. 49. Rawlins, “Give the Emperor a Mirror.” 50. Rawlins, “Give the Emperor a Mirror.” 51. Craft and Heim, “Transparency in Journalism,” 221. 52. W. Timothy Coombs and Sherry J. Holladay, “Communication and Attributions in a Crisis: An Experimental Study in Crisis Communication,” Journal of Public Relations Research 8 (November 1996): 279-95. 53. Coombs and Holladay, “Communication and Attributions in a Crisis.” 54. Rawlins “Give the Emperor a Mirror.” 55. David Good, “Individuals, Interpersonal Relations, and Trust,” in Trust: Making and Breaking Cooperative Relations, ed. Diego Gambetta (Oxford, UK: Department of Sociology, University of Oxford, 2000), 31-48; Hart et al., “Exploring Organizational Trust and Its Multiple Dimensions.” 56. Allen “The Trouble with Transparency.” 57. Rawlins “Give the Emperor a Mirror.” 58. Plaisance, “Transparency,” 197. 59. Allen “The Trouble with Transparency.” 60. Allen “The Trouble with Transparency,” 331. 61. Onora O’Neill, A Question of Trust (Cambridge, UK: Cambridge University Press, 2002). 62. Rawlins “Give the Emperor a Mirror.” 63. W. Timothy Coombs and Sherry J. Holladay, “Further Explorations of Post-crisis Communication: Effects of Media and Response Strategies on Perceptions and Intentions,” Public Relations Review 35 (March 2009): 1-6. 64. Stephens and Malone, “New Media for Crisis Communication.” 65. Sweetser “A Losing Strategy.” 66. Susan B. Walton, “Trendspotting: What You Can Expect in the Coming Year,” Public Relations Tactics (January 2010): 14-15. 67. Coombs and Holladay, “Further Explorations of Post-crisis Communication.” 68. Timothy W. Coombs and Sherry J. Holladay, “Helping Crisis Managers Protect Reputational Assets: Initial Tests of the Situational Crisis Communication Theory,” Management Communication Quarterly 16 (November 2002): 165-86. 69. Rawlins, “Give the Emperor a Mirror.” 70. Rawlins, “Measuring the Relationship.” 71. Good, “Individuals, Interpersonal Relations, and Trust.” 72. Good, “Individuals, Interpersonal Relations, and Trust,” 44. 73. Good, “Individuals, Interpersonal Relations, and Trust.” 74. McAllister, “Affect- and Cognition-Based Trust.” 75. McAllister, “Affect- and Cognition-Based Trust.” 76. Tom Kelleher, “Conversational Voice, Communicated Commitment, and Public Relations Outcomes in Interactive Online Communication,” Journal of Communication 59 (March 2009): 172-88; Tom Kelleher and Barbara Miller, “Organizational Blogs and the Human Voice: Relational Strategies and Relational Outcomes,” Journal of Computer-Mediated Communication 11 (August 2006): 395-414.

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014 Kim et al. 829

77. Alfonso Gonzalez-Herrero and Suzanne Smith, “Crisis Communications Management on the Web: How Internet Based Technologies Are Changing the Way Public Relations Professionals Handle Business Crises,” Journal of Contingencies and Crisis Management 16 (August 2008): 143-53. 78. Hart et al., “Exploring Organizational Trust and Its Multiple Dimensions”; McAllister, “Affect- and Cognition-Based Trust.” 79. Michael D. Basil, William J. Brown, and Mihai C. Bocarnea, “Differences in Univariate Values versus Multivariate Relationships,” Human Communication Research 28 (January 2002): 501-14. 80. Richard Whittington and Basak Yakis-Douglas, “Strategic Disclosure: Corporate Reputation and the Communication of Strategy,” (September 2010), http://www.sbs. ox.ac.uk/sites/default/files/CCR/Docs/2010Symposia-WP-YakisDouglasWhittington.pdf (accessed April 30, 2011). 81. Laura Arpan and Donnalyn Pompper, “Stormy Weather: Testing ‘Stealing Thunder’ as a Crisis Communication Strategy to Improve Communication Flow between Organizations and Journalists,” Public Relations Review 29 (3, 2003): 291-308; Laura M. Arpan and David R. Roskos-Ewoldsen, “Stealing Thunder: Analysis of the Effects of Proactive Disclosure of Crisis Information,” Public Relations Review 31 (September 2005): 425-33. 82. Mayer, Davis, and Schoorman, “An Integrative Model of Organizational Trust.”

Downloaded from jmq.sagepub.com at Istanbul Universitesi on December 20, 2014