Inclusive Discourse? Local Media Coverage of the Terri Schiavo Case

Abstract

This article contributes to a more systematic understanding of the role of mainstream media in deliberative processes by analyzing how inclusive local newspapers are of diverse perspectives on the Terri Schiavo case. Using both Pearson’s chi-square test and multinomial logistic regression, we assess how ownership, state political ideology, geographic location, and news format affect what ideas are included in the debate over whether Terri Schiavo’s hydration and nutrition tubes should be removed as well as the tone with which these ideas are discussed. We find that mainstream newspapers are relatively inclusive of diverse ideas and perspectives—regardless of whether the newspaper is independently or corporately owned, the political leanings of the target audience, and the geographic location of the outlet. However, we also find that local newspapers do significantly differ from one another in terms of the frequency and tone with which they include diverse viewpoints. Our research suggests that local outlets downplay ideas that are likely to be regarded as controversial by their target audiences. We conclude with a discussion of the implications of this research for understanding the role of media in deliberative democratic processes in the United States.

Communication is critical to creating and sustaining a robust democracy because it makes public interaction, discourse and, ultimately, coordinated action possible (Habermas 1984). Communicative action simultaneously involves promoting consensus among pluralistic groups (which prevents the dissolution of a democratic society) and fostering conflict by allowing groups to articulate their more particularistic positions. The inclusion of diverse viewpoints in public debate, then, is important to deliberative processes because it brings new ideas and “changes of the will” from the periphery to the center of a society (Dahlberg 2005; Habermas 1996). While differences in the perspectives are rarely resolved, the inclusion of diverse points of view helps individuals transcend their subjectivity, integrate other perspectives into their understanding of what constitutes a just and democratic society, and, ultimately, better equips them to carry out their civic duties (Porto 2007; Young 1997). In short, the inclusion of diverse viewpoints in the political conversation makes deliberative democratic processes more robust (Jacobs and Townsley 2011; Wessler 2008).

Social scientists, however, question the extent to which mainstream news media are “multiper-spectival,” or representative of diverse points of view (Gans 1979; Gans 2011), and, consequently, their ability to contribute to deliberative democratic processes meaningfully (Ferree et al. 2002; Sobieraj 2011).¹ For example, some scholars argue that professional norms systematically limit the range of ideas included in mainstream media coverage. In practice, the preoccupation with crafting “objective” news stories, means that journalists try to “balance” their stories by including contrasting points of view in coverage (Gamson et al. 1992; Tuchman 1972). This practice, however, does not necessarily translate into coverage that is inclusive of the range of positions on an issue or event (Sobieraj 2010; Wessler 2008). Other scholars assert that the changing structure of the news industry, and the emphasis on profitable rather than quality coverage, narrows the range of ideas included in the news (Baker 2002; Baker 2006; Cranberg, Bezanson, and Soloski 2001; Picard 1989; Picard and van Weezel 2008). As local news stations and newspapers are bought up by large corporations with several media holdings, the staff of regional outlets shrinks. Local journalists, in turn, rely more heavily on prepackaged news and institutional sources, which systematically narrows the range of ideas included in coverage across the country (Bagdikian 2004; Parenti 1993). This undermines the ability of mainstream news to contribute to deliberative democratic processes because diversity of opinion and point of view are sacrificed for efficiently produced, profitable news products (Baker 2002; McChesney 1999).

We argue that it is too soon to declare mainstream media outlets unrepresentative of diverse viewpoints because current research largely ignores local media coverage. Sociologists who systematically analyze mass media tend to do so with broad strokes and often make generalizations about the mainstream and its deleterious effects on discourse based on treatments of prominent, nationally circulated outlets, such as broadcast news and The New York Times, or “agenda setting” regional outlets like the Los Angeles Times (Benson 2013; Ferree et al. 2002; Gamson 1992; Gitlin 1980; Rojecki 1999).² While these news outlets are influential, they alone do not constitute mainstream media. Despite changes in the news industry, including technological ones, local outlets are a significant informational touchstone for the citizenry.³ More importantly, when scholars do analyze local outlets they find that the news processes and output are very different than those of national venues (Andrews and Caren 2010; Franklin and Murphy 1998; Kaniss 1997). In short, evidence suggests that it is too soon to convincingly argue that national or agenda-setting news outlets are adequate proxies for coverage at the local level—let alone to deliver a verdict on the inclusivity of mainstream news coverage.

This article contributes to a more systematic understanding of the role of mainstream media in deliberative processes by analyzing how inclusive local newspapers are of diverse perspectives on the Terri Schiavo case. The Schiavo case garnered national media attention when Terri’s parents launched a campaign to prevent the removal of Terri’s hydration and nutrition in 2000. Here, we triangulate our methods and use both Pearson’s chi-square test and multinomial logistic regression to assess how ownership, state political ideology, geographic location, and news format affect what ideas are included in the debate over whether Terri Schiavo’s hydration and nutrition tubes should be removed as well as the tone with which these ideas are discussed. We find that mainstream newspapers are relatively inclusive of diverse ideas and perspectives—regardless of whether the newspaper is independently or corporately owned, regardless of the political leanings of the target audience, and regardless of the geographic location of the outlet. That said, we also find that local newspapers do significantly differ from one another in terms of the frequency and tone with which they include diverse viewpoints. Our research suggests that local outlets downplay ideas that are likely to be regarded as controversial by their target audiences. We conclude with a discussion of the implications of this research for understanding the role of media in deliberative democratic processes in the United States.

LOCAL MEDIA AND THE POTENTIAL FOR INCLUSIVE DISCOURSE

Scholarly understandings of the factors that affect news media coverage vary according to their empirical aims. Communication scholars interested in how society influences coverage, for example, sometimes examine the ways in which the characteristics of a community affect what (and how) local news is constructed (e.g., Demers and Viswanath1999; Pollack 2007). While a community approach is valuable, sociologists typically utilize organizational or institutional approaches to understand what factors shape the news and how (Benson and Saguy 2005; Clayman and Reisner 1998; Gamson 1992; Gamson et al. 1992). Consistent with this sociological work, we conceptualize mass media as an institutional field and highlight the factors that potentially affect how inclusive local media outlets are of diverse points of view.

Currently, two complementary variants of field theory animate analyses of American news media—the organizational fields discussed by new institutionalists and field theory as outlined by Pierre Bourdieu.⁴ The two variants of field theory have much in common. Both conceptualize a field as a structured social space that is comprised of a network of relationships among actors with more or less power and both explain regularities in individual and organizational behavior by situating actors within a field of action. Likewise, both variants argue that fields are relatively coherent because actors are oriented toward a particular value or prize and agree on the “rules of the game” by which these values are accumulated (Benson 2006; Bourdieu 1998b; DiMaggio and Powell 1983; Martin 2003). The perspectives vary, however, in their understanding of the relational dynamics animating institutional fields and, ultimately, what this means for the potential of inclusive discourse.

New institutionalists emphasize the processes through which organizations in a field come to resemble one another, or institutional isomorphism (DiMaggio and Powell 1983). New institutionalists conceptualize an organizational field as “organizations that, in the aggregate, constitute a recognized area of institutional life: key suppliers, resource and product consumers, regulatory agencies, and other organizations that produce similar services or products” (DiMaggio and Powell 1983:148). Actors operating in a field are guided by an institutional logic, which consists of “a set of material practices and symbolic constructions” (Friedland and Alford 1991:248) that provide the organizing principles which outline how values (e.g., money, audience share, and prestige) are accumulated (Meyer and Rowan 1977). As a field becomes more homogenous, the rules, resources, and relational networks solidify and constrain the actions and choices of organizations operating in the field.⁵ In short, new institutionalism argues that, although organizations are strategic actors that may alter their practices and goals, the dominant norms and practices in a field homogenize organizational practices as well as circumscribe an organization’s activities and potential to change (Fligstein 2001; Hensmans 2003; Rao, Morrill, and Zald 2000).

New institutionalism, often implicitly, informs analyses of journalism as both an occupation and a commercial enterprise.⁶ Not surprisingly, this research highlights the aspects of the news production process that potentially homogenize coverage in commercial, mainstream news outlets. Sociologists, for instance, analyze how occupational professionalization and news production processes homogenize mainstream news coverage. Mainstream journalists are expected to be objective. As such, journalists are socialized to cover events and political happenings with fact-driven and politically neutral reporting (Schudson 2003; Sigelman 1973; Tuchman 1972).⁷ On the ground, journalists, who consistently deal with scare resources and tight deadlines, practice neutrality by simply including opposing points of views—as opposed to a range of viewpoints—in their stories (Gamson et al. 1992; Tuchman 1972). Likewise, journalists adopt routines that make finding and reporting the news of the day easier. Local journalists, for example, look to news services and prominent national media outlets such as The New York Times to determine what stories they should include in their own coverage (Gans 1979; Tuchman 1978b). Journalists also routinize their news beats to frequent locations where news typically occurs (e.g., police stations, court houses, and governmental buildings) and feature the actors best positioned to offer informed comments on the news of the day (Tuchman 1978a). These processes, scholars argue, create biases in reporting which homogenize what events and ideas are included in coverage (Amenta et al. 2009; McCarthy, McPhail, and Smith 1996; Smith, McCarthy, McPhail, and Augustyn 2001) and, ultimately, undermine democratic processes (Gans 2003; McChesney 1999).

While this work indeed is important, new institutionalism’s emphasis on homogeneity in the field assumes that the news production processes and stories circulated in local outlets mirror those of national and agenda-setting outlets. However, there is evidence to the contrary and, consequently, reason to suspect that local outlets may ultimately vary from one another in terms of whether and how they include diverse viewpoints. In short, scholars need to recognize the potential for differences in the field and, more specifically, focus their attention on the factors that make local coverage more (or less) inclusive. Bourdieu’s conceptualization of the journalistic field provides a useful starting point in this regard. Unlike new institutionalism, Bourdieu (1998a, 2005) emphasizes heterogeneity, or the processes through which outlets distinguish themselves from one another, in the journalistic field. He argues that the journalistic field is organized around two opposing forms of power—economic and cultural capital (Bourdieu 1998a, 1998b, 2005). In the journalistic field, economic capital is represented by circulation/ratings and advertising dollars and cultural capital by professional honors that result from peer recognition such as the Pulitzer Prize (Benson 2006; Bourdieu 1998a). The field is structured around two poles: the “heteronomous pole,” which represents forces external to the field (in this case, the market and assessments regarding outlet profitability), and the “autonomous pole,” which represents the specific form of capital valued within a field (in this case, intellectual reporting) (Benson 2006). Since a field is a structured space between these two poles, an outlet’s location within the field indicates the kinds of internal and external pressures that come to bear on the outlet as well as the amount of capital it wields (Bourdieu 1998a). Understanding the position of an outlet in the field, then, illuminates the pressures to which news venues respond, highlights how these pressures affect media coverage of national and international issues (Benson 2009; Benson and Hallin 2007; Benson and Neveu 2005; Benson and Saguy 2005), and sheds light on what these pressures potentially mean for the inclusivity of discourse and deliberative democratic processes more generally (Benson 2013; Rohlinger 2007).

Mainstream media outlets are located near the heteronomous pole, which means they are disproportionately influenced by economic considerations and the relational dynamics among these venues largely are competitive (Bourdieu 1998a, 2005). That said, because the field is hierarchically structured with news venues targeting different audiences, some outlets compete more directly than others. As a result, multiple news outlets can occupy a similar space in a field without competing with one another for audiences or advertising dollars (Benson 2013). This hierarchy creates heterogeneity in the field insofar economic pressures on outlets can exact different effects, which, in turn, may influence whether and how a given outlet includes a range of perspectives on an issue. In short, Bourdieu’s variant of field theory suggests that since local outlets do not compete with one another directly, we should not assume that they cover diverse ideas at the same rate or even with a similar tone. In fact, some local mainstream outlets may be more inclusive than others, giving some audiences a better opportunity to familiarize themselves with diverse perspectives on important social and political issues.

There are at least four factors, three of which represent economic considerations, that potentially shape the inclusivity of diverse viewpoints in local media coverage: the ideological orientation of the outlet’s target audience, the geographic location of the outlet, outlet ownership, and news format. Like national outlets, local news venues regard their audience as consumers, who they want to deliver to advertisers. Local outlets do so by reporting on events that are relevant to the desired demographic in ways that generally resonate with their political sensibilities (Cranberg, Bezanson, and Soloski 2001; Gartner 2004; Kaniss 1997). As a result, local outlets may discuss perspectives that resonate with their target audience at higher rates or, alternatively, avoid those ideas that may anger their desired demographics and potentially lead to a loss of revenue. This, in turn, may affect the inclusivity of discourse in terms of the frequency with which issues and events are discussed as well as how they are presented to target audiences.

Similarly, the geographic location of an outlet may affect how it covers an issue or event. The closer an outlet is to the epicenter of an event the more in-depth coverage it receives (Gartner 2004). The sheer amount of coverage may have implications for deliberative democratic discourse. As journalists work to maintain the audience’s attention, they look for new ways to cover ongoing issues and events (Gans 1979; Tuchman 1973). This, in turn, may lead to more inclusive coverage and improve the range of ideas to which an audience is exposed. However, the relative inclusivity of discourse may decline the further away a local outlet is from the issue or event, particularly if one or more perspectives conflict with the political sensibilities of an outlet’s desired demographic.

In addition to economic factors, the “news format” (Cottle 1995) may also affect the range of ideas covered by a local outlet. Target audiences are not monolithic in their perspectives on social and political issues. One way for local outlets to diversify the ideas discussed is to include them in opinion formats such as political cartoons, editorials, and letters to the editor (Gamson and Stuart 1992; Wahl-Jorgensen 2002). While this may seem like an obvious point, scholars typically focus on news coverage in their analyses of discourse and deliberative processes (except see Benson 2013; Jacobs and Townley 2011). The emphasis on news alone is misplaced because opinion formats provide an important free space for the discussion of diverse ideas, meaning they have the potential to make discourse more inclusive by discussing ideas that are generally excluded from the news (Jacobs and Townsley 2011).

DATA AND METHODS

We use data derived from a content analysis to analyze the inclusiveness of local coverage in general audience newspapers. Here, we assess the presence of diverse views in coverage of the Terri Schiavo custody case as well as the tone of the coverage (discussed below). Terri Schiavo, who collapsed in 1990 as the result of unknown causes, did not become a touchstone for national debate until 2000; the year a Pinellas County court granted her husband, Michael Schiavo, permission to withdraw Terri’s hydration and nutrition tubes. Terri’s parents, Robert and Mary Schindler, opposed the decision and turned to the political system for recourse. With the help of pro-life leaders, Robert and Mary Schindler organized a vigil outside of Terri’s hospice and presented the governor of Florida (Jeb Bush) a petition with 180,000 signatures asking him to prevent Schiavo’s death. The Florida legislature responded and passed a bill allowing Bush to reinstate Terri’s hydration and nutrition; a move ultimately deemed unconstitutional by the Florida State Supreme Court. The scope of the political conflict widened when the U.S. Congress became involved and passed a relief bill that transferred jurisdiction of the Schiavo case to the federal courts; a bill which President George W. Bush signed into law. The Schindlers immediately requested an emergency injunction in the U.S. District Court and asked for the reinstatement of Terri’s hydration and nutrition. The request was denied. The U.S. Supreme Court declined to hear the case and to grant certiorari. Terri Schiavo passed away a few days later on March 31, 2005.

The Terri Schiavo case is ideal for examining the inclusion of diverse ideas and perspectives in local mainstream coverage. First, the Schiavo case involved a controversial issue (whether her hydration and nutrition should be removed) and produced a firestorm of national media coverage. Consequently, hundreds of U.S. newspapers carried at least one story or opinion piece referencing the debate over whether Terri’s hydration and nutrition tubes should be removed. In short, the Schiavo case was a salient one, which affords an excellent opportunity to analyze inclusivity systematically across local outlets. Second, actors, outside of Terri’s immediate family, were interested in the case (e.g., politicians, medical professionals, ethicists, lawyers, and disability activists). These diverse actors used a broad range of arguments to advocate for and against the removal Schiavo’s hydration and nutrition as well as to talk about social issues more generally. Some actors, for example, focused on the specifics of Terri’s life and death, while others emphasized the role of politics in individual decision making, the importance of living wills, the treatment of persons with disabilities in American society, or the flaws in medical ethics and the medical system. Given the diversity of argumentation championed, the Schiavo case provides a good opportunity to assess the extent to which these very different arguments are included in coverage as well as whether factors such as the political leanings of the target audience, the proximity of the outlet to the event epicenter, and news format affect when ideas are included and how. Finally, media owners did not have an economic stake in how the case played out, which means coverage does not simply reflect owner impulses to protect their (and their advertisers) economic interests (Croteau and Hoynes 1994). This, in turn, allows us to more accurately consider whether there are differences in what ideas local newspapers include in coverage apart from a corporation’s (or owner’s) other holdings.

Dependent Variables

The dependent variables are the frames used to advocate for or against the removal of Terri Schiavo’s hydration and nutrition. In order to derive a list of frames that represented a broad range of positions that might be excluded or included in coverage, the lead researcher began by analyzing the documents, press releases, and reports of 16 organizations with differing perspectives on the Terri Schiavo case. The groups included disability rights organizations (ADAPT, Advancing Independence: Modernizing Medicare and Medicaid, Disability Rights Center, and Not Dead Yet); a death with dignity group (Compassion & Choices, which is the same as End of Life Choices and Compassion in Dying aka the Hemlock Society); senior organizations (AARP and Aging with Dignity), a legal organization (American Bar Association), medical organizations (American Medical Association and the National Hospice Foundation), a pro-life organization (National Right to Life Committee), conservative organizations (Christian Coalition, Heritage Foundation, and the Family Research Council), a religious organization (the National Catholic Partnership on Disability) and a foundation started by the Schindlers (Terrisfight.org).⁸ Drawing on this material, the lead researcher constructed a list of 42 specific arguments used to support six general frames: The legal, medical, political, disability, right to life, and death with dignity frames.⁹ Table 1 summarizes each frame and provides examples of the more specific arguments used to support this general interpretation of the issue.

TABLE 1.

Summary of Frames

Frame (with examples)	General Description of the Frame
Legal Frame	Discusses the implications of the Schiavo case to individual rights and/or the legal system.
Right to privacy	Medical decisions by individuals and family members are private and constitutionally protected.
Judicial independence	The legislative and executive branches of government should not interfere with judicial affairs or decision-making.
Medical Frame	Discusses Terri’s diagnosis, condition, and/or treatments.
Withholding valuable treatment	Treatments that would improve Terri’s cognitive functions/quality of life are being intentionally withheld.
Disputed diagnosis	There is not a consensus regarding the diagnosis that Terri is in a “persistent vegetative state,” nor is their agreement on what this diagnosis means.
Political Frame	Discusses political issues and interventions or their implications for government, governing, or political parties.
Legislative abuse	Politicians are abusing their authority and power in order to promote their own agenda.
Flawed legislation	Legislative interference increases the reach/power of the government in individual decision making.
Death with Dignity Frame	Argues that terminally ill, mentally competent adults have the right to choose the time, place, and circumstances of their death.
Terri’s wishes	Terri expressly stated that she did not want to live with machines or in a state of extreme disability.
Individual choice	Artificial nutrition/hydration is medical treatment that individuals can choose to accept or reject.
Disability Frame	Discusses the rights and interests of persons with disability or how persons with disabilities are treated in the U.S.
Deadly discrimination	The current legal/political/medical situation allows disabled people to be killed because of their health status.
Right to Life Frame	Argues that society has a moral obligation to intervene and protect life from natural birth to natural death.
Consequences of humanism	The rise in humanism and moral relativism in the U.S. diminishes our respect for life.

We coded for the presence of all general frames and specific arguments in the media stories. Since frames could appear multiple times in a single article, we coded each time a frame was mentioned in a case. We also coded whether the speaker championing the argument used partisan (e.g., advocated for or against a position) or neutral (e.g., simply presented an argument) language. This is an important step because while an actor or journalist may discuss a particular idea in coverage, s/he does not necessary do so neutrally. For instance, an actor may passionately disagree that Terri Schiavo’s diagnosis is disputed among medical professionals and, then, argue that this conflict is a ruse designed to withhold valuable treatment (arguments associated with the medical frame). A frame was coded “partisan” if it was clear from the language used that an actor agreed or disagreed with the argument (e.g., “It is obvious to any educated citizen that Terri is being discriminated against”). A frame was coded as “neutral” if an actor mentioned or discussed the argument but did not indicate support for or opposition to the ideas (e.g., “Some disability groups argue that Terri is being discriminated against by the medical establishment”). It is worth noting that some of these frames are likely to get included in coverage more often than others. Journalists prefer to include ideas that resonate with institutional and cultural frameworks over those that do not (Ferree 2003). This means that the legal, medical, and political frames, which align with established institutions, may get included in coverage more often than frames that do not (e.g., the death with dignity, disability, and right to life frames).

There is an additional point worth making about article tone. While we initially coded whether a tone was “supportive” or “oppositional,” small cell counts for some of the frames prohibited us from keeping the distinction here. In an effort to more accurately capture the tone of these rarely mentioned frames, we assigned each frame an “overall” tone score at the story level (which is the unit of analysis). For example, if a frame was discussed using a neutral tone in coverage, than it was assigned as neutral at the story level. If a frame appeared three times in the same story, however, and was discussed using a neutral tone once and a partisan tone twice, it was coded as partisan overall. This decision also made sense given our use of multinomial logistic regression, which as we discuss below, can be difficult to interpret when there are multiple outcomes.

This study, which examines media discourse in U.S. local newspaper coverage between 2000 and 2004, includes a total of 1,182 local newspaper stories. Of these, 774 are news stories and 408 are opinion pieces and letters to the editor—less than 10% of the sample are news wire stories or syndicated opinion pieces. The sample was generated using Lexis Nexis Academic Universe and the search term “Terri Schiavo.” When creating the sample we threw out articles that mentioned, but did not discuss, the Terri Schiavo case and duplicate articles (or articles from the same date with the same word count from the same outlet that were captured twice by the search engine).¹⁰ Additionally, our sampling strategy was designed to capture potential shifts in discourse. Given the huge number of articles written about the case and the limited resources available to conduct the research, the lead researcher ultimately decided to code coverage during the period in which the most critical events in the case occurred. To ensure that media coverage did not fundamentally change in the last months of the debate, a random sample of 10% of the media coverage was coded and compared to the data presented here. There were not significant differences between the samples.¹¹

The media stories were coded between February 2009 and February 2010 by the lead author and a team of graduate students. To ensure intercoder reliability, a project manager checked coding consistency among team members weekly. Reliability was assessed two ways. First, the project manager randomly selected and coded at least 10% of the weekly coding output for each team member. The project manager would, then, assign a consistency score based on the percent of divergent codes between project manager and individual team member’s data. Second, a subset of randomly chosen articles was coded by four team members simultaneously on a weekly basis. This allowed the project manager to check the reliability across coders and derive a reliability score for each coder.

$$\text{Realibility Score}=\frac{\text{Matched Codes}}{\text{Unmatched}+\text{Matched Codes}}$$

Team members with scores lower than 90% on either reliability check attended an additional training session and the emails were redistributed and recoded. While this slowed down the coding process considerably, this strategy was very effective yielding an overall reliability score of 92.7%.

Independent Variables

In order to assess whether an outlet’s position in the field affects the inclusivity of diverse viewpoints in local coverage, we include four independent variables in our analysis: media ownership, the political ideology of the state in which an outlet is located, geographic location of the outlet, and news format. First, we examine whether corporate ownership—specifically, whether the outlet was independently or corporately owned during the 2000–2004 time frame, affects the inclusion of the different frames. Stories and editorials about the Terri Schiavo case appeared in 139 different newspapers across the United States. Since past ownership data for regional and local outlets is not readily available, we searched each outlet’s website in order to assess the status of the newspaper between 2000 and 2004. An outlet was considered corporately owned if its parent company held more than one local or national newspaper.¹²

Second, we include a political ideology variable in order to assess whether the political leanings of a state affect local coverage of the Schiavo case. Recall that the existing research on local media coverage finds that outlets tailor content to the sensibilities of their target audiences (Cranberg et al. 2001; Kaniss 1997). Thus, it is reasonable to expect that outlets serving more politically liberal demographics to cover controversial issues differently than outlets targeting conservative ones. For example, newspapers in Washington and Oregon, where death with dignity laws have passed, may cover the Schiavo case differently than those in Louisiana, where the citizenry is far more conservative on “right to life” issues. Each state was categorized as conservative, moderate, or liberal using the following two steps. First, we classified a state using Medoff’s (1997) operationalization of state political ideology. Medoff determines the political ideology of a state through an analysis of U.S. Congressional representatives’ voting histories. He contends that this is a good way to determine how liberal, moderate, or conservative a state is because representatives are expected to vote the preferences of their constituents (also see Roh and Berry 2008). While other measures of state political ideology are more widely used (see Berry et al. 1998, 2010) the Medoff classification is the most useful for our purposes because it categorizes states into three distinct groups rather than plotting states on a continuous scale ranging from 0–100. While using categorical measures of state ideology does not allow the precision that the Berry et al. scale provides, it better fits the statistical modeling used to analyze this data.¹³

Since Medoff’s measure does not necessarily work well for classifying states on social issues, which may appear on state ballots far more frequently than in Congress, we also considered the abortion policies of each state and their presumed willingness to liberalize or restrict abortion access if the Supreme Court overturns its 1973 Roe v. Wade decision.¹⁴ To do so, we drew on a 2006 report in USA Today that categorized states as (1) likely to protect abortion access, (2) likely to restrict abortion access, or (3) moderate on abortion and unlikely to make dramatic laws one way or the other (Page 2006). These categories, we reasoned, corresponded with liberal, conservative, and moderate states respectively. This strategy for categorizing states worked well in most cases. For instance, states that Medoff identified as conservative often had restrictive abortion policies and were considered likely to restrict abortion access further if the opportunity presented itself. However, if a state ranked differently on these measures, we used the results of the 2000 and 2004 presidential election, which are the elections most closely related to our time frame, as a tie breaker. New York, for instance, has a moderate state political ideology (Medoff 1997) and liberal abortion laws (Page 2006). Since the state went to the Democratic candidate in 2000 and 2004, we categorized it as a “liberal” state on social issues. Similarly, Wisconsin ranks as “conservative” on the political ideology scale, but is relatively liberal in its abortion access policies. Since the state went to the Democratic candidate in 2000 and 2004, we categorized it as “moderate” on social issues. In short, this approach to categorizing state political ideology helps capture some, although not all, of the diversity of the citizenry and, consequently, provides a better measure of whether the political sensibilities of the citizenry affect whether and how ideas are discussed in local newspaper coverage.¹⁵

Third, since states close to the epicenter of an event may do more stories on an issue and be more inclusive in the kinds of frames discussed in coverage than those further away, we include a geographic region variable to assess whether the frequency and tone of coverage varies according to a state’s relative distance from Florida. For the purposes of this study we used the Census Bureau’s classification to categorize newspapers into one of four regions: West (13 states), Midwest (12 states), Northeast (9 states), and South (17 states).¹⁶

Finally, because the opinion pages have the potential to diversify what and how ideas regarding the Schiavo case are covered, we include a measure of news format. Each article was categorized as either a news story or an editorial. The latter included letters to the editor and opinion pieces (e.g., anything that appeared on the opinion page as well as syndicated columns). Generally speaking, editorials advocated for a specific position or course of action on the Terri Schiavo case (or castigated a position articulated by a writer or journalist in the paper). Articles that were designed to inform or update the readership on the Schiavo debate were categorized as news stories.

We triangulate our methods in order to analyze how inclusive local outlets are of diverse ideas on the Schiavo case. First, we use Pearson’s chi-square test to assess whether the distributions of the frames (specifically the frequency with which frames are excluded from coverage, included in coverage using a neutral tone, and included in coverage using a partisan tone) varies by ownership, political ideology, geographic region, and news format. Then, we use multinomial logistic regression to assess the effects of the four independent variables—ownership, political ideology, geographic location, and article type—on the presence and tone of the six frames. Due to the nature of the dependent variable, which is categorically distributed—not mentioned, neutral tone (referent), and partisan tone—the multinomial logistic regression model is most appropriate for our analyses. We report the relative risk ratios (exponentiated coefficients) and standard errors in parentheses.

FINDINGS

Table 2 shows whether each frame is not mentioned in coverage, included in coverage using a neutral tone, or included in coverage using a partisan tone in corporately and independently owned newspapers. Recall that the Pearson’s chi-square test indicates whether or not there are significant differences in the distribution of the frames by each of the independent variables. This test provides a useful starting point because it shows the differential effect of an independent variable on the dependent variables. In general, corporately and independently owned newspapers include (and do not mention) frames at similar proportions and in similar ways. Local newspapers, regardless of ownership, regularly include the legal, medical, political, and death with dignity frames in coverage. Similarly, corporately and independently owned newspapers limit their discussion of the disability and right to life frames relative to the Terri Schiavo case, as indicated by the relatively high proportion of “not mentioned” in each of these frames. While the relative visibility of institutional frames (legal, medical, and political frames) and obscurity more particularistic frameworks (disability and right to life frames) is consistent with the current literature (Ferree 2003), the relatively frequent inclusion of the death with dignity frame is somewhat surprising since it is a highly contested practice in the United States. That said, it is possible that journalists saw the frame as more relevant—and relatively less controversial—to understanding the Schiavo than the disability and right to life perspectives.

TABLE 2.

Presence of the Frames by Ownership

	Corporately Owned %	Independently Owned %
Legal Frame
Not Mentioned	21.4	20.0
Neutral Tone	51.6	49.4
Partisan Tone	27.0	30.6
Total	100.0	100.0
Medical Frame+
Not Mentioned	31.0	22.4
Neutral Tone	54.5	55.3
Partisan Tone	14.5	22.4
Total	100.0	100.0
Political Frame**
Not Mentioned	46.8	29.4
Neutral Tone	32.0	45.9
Partisan Tone	21.2	24.7
Total	100.0	100.0
Disability Frame
Not Mentioned	79.5	70.6
Neutral Tone	10.0	14.1
Partisan Tone	10.5	15.3
Total	100.0	100.0
Right to Life Frame
Not Mentioned	88.8	87.1
Neutral Tone	7.6	9.4
Partisan Tone	3.6	3.5
Total	100.0	100.0
Death with Dignity Frame
Not Mentioned	43.8	34.1
Neutral Tone	26.3	29.4
Partisan Tone	29.9	36.5
Total	100.0	100.0
Total N	1,097	85

Note. Column statistic presented. Due to rounding columns may not equal 100%.

⁺p ≤ .10;

^*p ≤ .05;

^**p ≤ .01;

^***p ≤ .001.

While the discourse of corporately and independently owned outlets both seem inclusive, the presence and tone of the frames vary significantly by newspaper ownership for the political frame (χ² = 10.28, p ≤ .01) and the medical frame (χ² = 5.18, p ≤ .10). Independently owned newspapers are more likely to discuss the political and medical frames as well as express support/opposition to these frames than corporately owned outlets. This is an important finding because it suggests that independently owned present more lively, and potentially more critical, perspectives on the purpose and practices of American institutions. For example, Terri’s persistent vegetative state diagnosis was supported and contested by doctors, scholars, and pundits alike. The fact that independently owned venues would present passionate perspectives that question institutional practices and highlight disputes within medical field suggests that these outlets indeed may do a better job of educating the public on some issues. Ownership, in short, seems to have little effect on the inclusivity of mainstream media coverage, but does influence how some ideas are discussed in coverage.

Table 3 shows how often each frame is included (or excluded) from coverage as well as the tone with which it was discussed by political ideology. While there are not significant differences in the distribution of the political, right to life, and death with dignity frames by political ideology, the chi-square tests indicate that the presence and tone with which the legal, medical, and disability frame were discussed vary between conservative, moderate, and liberal states. Newspapers in conservative and moderate states generally are more likely to include the legal (χ² = 16.17, p ≤ .01) and disability (χ² = 8.46, p ≤ .10) frames in coverage. That said, moderate states discuss a legal frame using a partisan tone more often than either conservative or liberal states. Additionally, although newspapers in different kinds of states include and exclude the medical frame in similar ways overall, discussion of this frame in conservative states is more partisan than that of moderate and liberal states (χ² = 9.14, p ≤ .10).

TABLE 3.

Presence of the Frames by Political Ideology

	Conservative %	Moderate %	Liberal %
Legal Frame**
Not Mentioned	19.6	20.9	28.6
Neutral Tone	53.9	42.5	48.2
Partisan Tone	26.5	36.6	23.1
Total	100.0	100.0	100.0
Medical Frame+
Not Mentioned	30.7	34.0	26.1
Neutral Tone	52.7	56.2	61.3
Partisan Tone	16.6	9.8	12.6
Total	100.0	100.0	100.0
Political Frame
Not Mentioned	46.5	42.5	43.7
Neutral Tone	30.7	38.6	38.2
Partisan Tone	22.8	19.0	18.1
Total	100.0	100.0	100.0
Disability Frame+
Not Mentioned	77.2	81.1	83.9
Neutral Tone	11.9	7.8	5.5
Partisan Tone	10.8	11.1	10.6
Total	100.0	100.0	100.0
Right to Life Frame
Not Mentioned	89.5	86.9	86.4
Neutral Tone	7.0	8.5	10.1
Partisan Tone	3.5	4.6	3.5
Total	100.0	100.0	100.0
Death with Dignity Frame
Not Mentioned	44.3	43.1	37.7
Neutral Tone	25.7	27.5	29.7
Partisan Tone	30.0	29.4	32.7
Total	100.0	100.0	100.0
Total N	830	153	199

Note. Column statistic presented. Due to rounding columns may not equal 100%.

⁺p ≤ .10;

^*p ≤ .05;

^**p ≤ .01;

^***p < .001.

These results are interesting because they suggest that local outlets may emphasize ideas that are relatively less controversial among their target audiences. For example, outlets in conservative states do not significantly differ from liberal and moderate states in terms of the presence and tone of the political frame but do on the medical frame, even though the former is potentially more relevant to their target audience’s working knowledge of democracy. The Schiavo situation pitted social conservatives, who advocated for Terri’s right to life, against fiscal conservatives, who wanted the government to stay out of case. Inclusive discourse regarding the role of government in individual decision-making, then, would be particularly important to a robust deliberative process within states where this is not a point of agreement. Instead, these results suggest that local newspapers in conservative states may have tried to downplay this ideological divide, and avoid alienating readers, by emphasizing the debate over Terri’s treatment and diagnosis in coverage. Outlets in conservative states are not unique in this regard. Similarly, outlets in moderate states may have avoided discussions regarding the individual and legal rights of Schiavo in an effort to cover the issue without angering their target demographics, who may fundamentally disagree on who should get to determine Terri’s fate. Local outlets, in short, may be inclusive, but still may shy away from discussions that are the most likely to inflame their target audiences.

Significant differences also appear when exploring the relationship between frames and geographic region (Table 4). It is worth noting that the number of stories written about the Terri Schiavo case varies geographically. Newspapers located in and around Florida (represented by the category Conservative in Table 3 and South in Table 4) did more stories and editorials on the Schiavo case than newspapers in other parts of the country. In fact, there were almost five and one-half times more stories about the Schiavo case in Southern states than there were in Western states (Table 4). That said, newspapers in the South are not necessarily more inclusive of diverse perspectives than other regions of the country. For example, Northern states are the most likely to mention the disability (χ² = 12.25, p ≤ .10) and right to life frames (χ² = 22.21, p ≤ .001) in their outlets compared to other regions of the United States—and to discuss these frames using a partisan tone. Similarly, outlets in the West, North, and Midwest include discussions regarding the role of government in the Schiavo case (political frame) more often those in the South (χ² = 15.36, p ≤ .05). These findings suggest that inclusivity is not a function of the political leanings of an outlet’s target audience alone. The geographic location of an outlet seems to affect its willingness to deal with controversial ideas head on.

TABLE 4.

Presence of the Frames by Geographic Region

	West %	North %	Midwest %	South %
Legal Frame
Not Mentioned	24.5	27.3	19.9	19.6
Neutral Tone	49.6	50.6	47.3	52.8
Partisan Tone	25.9	22.1	32.9	27.6
Total	100.0	100.0	100.0	100.0
Medical Frame
Not Mentioned	28.1	25.0	30.1	32.1
Neutral Tone	59.7	61.1	56.9	51.6
Partisan Tone	12.2	14.0	13.0	16.3
Total	100.0	100.0	100.0	100.0
Political Frame*
Not Mentioned	44.6	44.2	39.0	47.3
Neutral Tone	41.0	33.7	41.8	29.5
Partisan Tone	14.4	22.1	19.1	23.2
Total	100.0	100.0	100.0	100.0
Disability Frame+
Not Mentioned	85.6	75.0	83.6	77.5
Neutral Tone	4.3	14.0	6.2	11.5
Partisan Tone	10.1	11.1	10.3	11.0
Total	100.0	100.0	100.0	100.0
Right to Life Frame***
Not Mentioned	92.8	79.1	87.0	90.5
Neutral Tone	5.8	14.5	9.6	6.1
Partisan Tone	1.4	6.4	3.4	3.5
Total	100.0	100.0	100.0	100.0
Death with Dignity Frame
Not Mentioned	43.9	37.8	41.8	44.4
Neutral Tone	27.3	33.7	26.0	24.8
Partisan Tone	28.8	28.5	32.2	30.8
Total	100.0	100.0	100.0	100.0
Total N	139	172	146	725

Note. Column statistic presented. Due to rounding columns may not equal 100%.

⁺p ≤ .10;

^*p ≤ .05;

^**p ≤ .01;

^***p ≤ .001.

The most dramatic differences in the distribution of the frames by tone emerge between news format and editorials. Table 5 shows that there are significant differences between news and opinion pieces in terms of whether or not (and how often) the legal (χ² = 93.88, p ≤ .001), medical (χ² = 57.86, p ≤ .001), political (χ² =18.28, p ≤ .001), and disability frames (χ² =11.87, p ≤ .01) are included in coverage. Generally speaking, these frames appeared less often in editorials, but were discussed using a more partisan tone. The differences by format are to be expected given that opinion pieces and Letters to the Editor are typically short in length and focus on a single argument, rather than trying to balance diverse points of view. That said, these results suggest that opinion pieces do enhance the inclusivity of local outlets by exposing readers to more diverse points of view (Jacobs and Townsley 2011). The medical frame, however, is an exception in this regard. The medical frame is discussed using a partisan tone more often in news stories than it is in editorials. This indicates that our argument that local outlets channel ideas that may upset target demographics to the opinion pages may be incorrect. We return to this point in the multinomial logistic regression analysis.

TABLE 5.

Presence of the Frames by News Format

	Article %	Editorial %
Legal Frame***
Not Mentioned	15.6	32.1
Neutral Tone	61.5	32.4
Partisan Tone	22.9	35.5
Total	100.0	100.0
Medical Frame***
Not Mentioned	23.1	44.1
Neutral Tone	59.2	45.8
Partisan Tone	17.7	10.1
Total	100.0	100.0
Political Frame***
Not Mentioned	45.0	46.6
Neutral Tone	36.6	26.2
Partisan Tone	18.5	27.2
Total	100.0	100.0
Disability Frame**
Not Mentioned	79.7	77.2
Neutral Tone	11.5	8.1
Partisan Tone	8.8	14.7
Total	100.0	100.0
Right to Life Frame
Not Mentioned	90.1	86.0
Neutral Tone	6.6	9.8
Partisan Tone	3.4	4.2
Total	100.0	100.0
Death with Dignity Frame
Not Mentioned	42.5	44.1
Neutral Tone	26.0	27.7
Partisan Position	31.5	28.2
Total	100.0	100.0
Total N	774	408

Note. Column statistic presented. Due to rounding columns may not equal 100%.

⁺p ≤ .10;

^*p ≤ .05;

^**p ≤ .01;

^***p ≤ .001.

Table 6 presents results from multinomial logistic regression analyses that show the effects of the four independent variables—ownership, political ideology, geographic region, and article type—on the coverage of the six frames. Here, we compute the relative risk ratios of the independent variables falling into three categories: (1) Not Mentioned, (2) Partisan Tone, and (3) Neutral Tone, which serves as the reference category.¹⁷ The relative risk is used to examine the “risk” of an outcome compared to a referent when there is another outcome in the model. For example, we assess the risk of independently owned outlets not mentioning the legal frame relative to the risk of corporately owned outlets presenting the same frame using a neutral tone (the referent). Interpreting a multinomial logistic regression model, as a result, can be complicated because it includes multiple outcome categories, which include a base category for comparison, as well as a set of independent variables with reference groups. A number greater than one indicates a positive association (or a higher odds) between the independent and dependent variable, while a number between zero and one indicates a negative association (or a lower odds). If the odds ratio equals 1.00, this indicates the odds are the same for the groups (Hoffmann 2004). Using the above example, then, we see that compared to corporately owned outlets, independently owned newspapers have greater odds of discussing the medical frame using a partisan tone, relative to discussing this frame using a neutral tone.

TABLE 6.

Relative Risk Ratios and Standard Errors from Multinomial Logistic Regression

	Legal Frame		Medical Frame		Political Frame
	Not	Partisan	Not	Partisan	Not	Partisan
	Mentioned	Tone	Mentioned	Tone	Mentioned	Tone
	(ref = Neutral Tone)		(ref = Neutral Tone)		(ref = Neutral Tone)
Ownership (Ref = Corporate)
Independent	0.95 (0.31)	1.17 (0.34)	0.71 (0.21)	1.88 (0.59)*	0.47 (0.13)**	1.02 (0.31)
Political Ideology (Ref= Moderate)
Conservative	0.66 (0.24)	0.54 (0.17)*	0.94 (0.29)	2.15 (0.90)+	1.00 (0.30)	1.25 (0.46)
Liberal	0.92 (0.36)	0.51 (0.18)+	0.67 (0.23)	1.89 (0.86)	0.81 (0.27)	1.09 (0.45)
Region (Ref=South)
West	1.05 (0.35)	1.02 (0.33)	1.01 (0.31)	0.59 (0.23)	0.91 (0.27)	0.48 (0.18)*
Midwest	0.82 (0.31)	0.85 (0.28)	0.85 (0.28)	1.15 (0.47)	0.63 (0.20)	0.69 (0.26)
North	1.24 (0.31)	0.85 (0.21)	0.72 (0.17)	0.71 (0.20)	0.92 (0.20)	0.85 (0.23)
News Format (Ref=Article)
Editorial	3.79 (0.61)***	2.94 (0.44)***	2.53 (0.35)***	0.75 (0.15)	1.47 (0.22)**	2.10 (0.36)***
	Disability Frame		Right to Life Frame		Death with Dignity Frame
	Not		Not		Not
	Mentioned	Partisan Tone	Mentioned	Partisan Tone	Mentioned	Partisan Tone
	(ref=Neutral Tone)		(ref=Neutral Tone)		(ref = Neutral Tone)
Ownership (Ref = Corporate)
Independent	0.46 (0.17)*	0.85 (0.38)	0.83 (0.34)	0.89 (0.64)	0.66 (0.20)	1.25 (0.37)
Political Ideology (Ref= Moderate)
Conservative	1.10 (0.49)	0.97 (0.55)	0.97 (0.49)	0.42 (0.33)	1.26 (0.41)	1.26 (0.44)
Liberal	2.00 (1.12)	1.79 (1.25)	0.83 (0.44)	0.43 (0.38)	0.87 (0.32)	1.62 (0.62)
Region (Ref=South)
West	2.58 (1.31)+	1.85 (1.14)	1.26 (0.62)	0.45 (0.43)	1.28 (0.41)	0.67 (0.23)
Midwest	2.18 (1.10)	1.62 (1.02)	0.62 (0.31)	0.33 (0.28)	1.15 (0.40)	1.12 (0.40)
North	0.68 (0.19)	0.65 (0.26)	0.40 (0.12)**	0.83 (0.43)	0.73 (0.17)	0.59 (0.15)*
News Format (Ref=Article)
Editorial	1.36 (0.30)	2.34 (0.64)**	0.66 (0.15)+	0.82 (0.31)	1.00 (0.15)	0.84 (0.14)

Note. The reference is “neutral tone.”

⁺p < .10;

^*p < .05;

^**p < .01;

^***p < .001 (two-tailed test).

The results of the multinomial logistic regression provide a more nuanced picture of how the independent variables affect the presence and tone of local media coverage. In contrast to Table 2, Table 6 shows that ownership does shape the frequency and tone with which outlets include some frames. Independently owned newspapers, for example, are significantly more likely than corporately owned outlets to mention the political and disability frames using a neutral tone, relative to having no mention of these frames in their stories.¹⁸ In addition, mentioning the medical frame using a partisan tone, relative to a neutral tone, is significantly more likely in independently owned newspapers when compared to corporately owned newspapers. These findings are important because they suggest that ownership matters insofar as independently owned outlets often present ideas, particularly ones that are controversial, more often than their corporately owned counterparts. Audiences of independently owned outlets, in other words, are exposed to debates about what role government should play in life-preserving and life-ending decisions, the extent to which the Schiavo case represented a disability rights issue, and passionate positions on the controversy regarding Schiavo’s diagnosis and treatment more often than those of corporately owned newspapers.

Consistent with the findings reported in Table 3, political ideology exerts few significant effects on the inclusion of diverse ideas (Table 6). That said, where differences do exist, political ideology significantly affects the tone with which ideas are presented. Compared to moderate states, conservative and liberal states are significantly less likely to mention the legal frame using a partisan tone (as opposed to a neutral tone) in their stories. On the other hand, conservative states are over two times more likely than moderate states to present the medical frame in their newspapers using a partisan tone, relative to a neutral tone. We did look for interaction effects between political ideology and ownership and only found one.¹⁹ This somewhat buttresses our claim that local outlets tailor their content to resonate with the political sensibilities of their target audience, although the practice is not necessarily wide spread. Newspapers serving audiences in moderate states focus on relatively safe debates regarding individual and legal rights, while liberal and conservative venues downplay them (comparatively speaking). Similarly, conservative outlets emphasize medical debates over frames that might divide their target demographic.

There are a number of significant differences when we consider the region in which the story appeared. For example, the South (where the Schiavo case occurred) and the West are significantly different from one another in terms of how they discuss the political frame and their inclusion of the disability frame. Compared to the South, outlets in the West are less likely to mention the political frame using a partisan tone, relative to a neutral tone, yet are over two times more likely to have not mentioned the disability frame (as opposed to mentioning the disability frame using a neutral tone). Moreover, outlets in Northern states are more likely than their Southern counterparts to mention the right to life frame, but less likely to mention the death with dignity frame using a partisan tone, relative to using a neutral tone. These findings generally support the argument that outlets closer to the epicenter of an event are more inclusive in terms of the presence of (and how) they cover a range of ideas. Here, the further away the outlet is from Florida, and the South more generally, the less likely it is to include sometimes passionate discourse on the rights of the disabled, the role of government in the Schiavo case, and whether individuals have a right to die with dignity.

Finally, as shown in Table 5, news format is significantly associated all of the frames, with the exception of the death with dignity frame. Generally speaking, editorials are significantly more likely to exclude frames from coverage than news stories. Again, this is expected given that opinion pieces and Letters to the Editor are typically short in length and focus on a single argument, rather than trying to balance diverse points of view. However, when compared to news articles, editorials also are significantly more likely to use a partisan tone (as opposed to a neutral tone), particularly when discussing legal, political, and disability frames. Again, passionate discussion regarding the role of politics and the implications of government involvement in the Schiavo case is not surprising. The push to intervene on Schiavo’s behalf (by Governor Jeb Bush, the largely conservative state legislator, and President George W. Bush) created conflict within the Republican Party as politicians argued over whether getting involved constituted an abuse of power and a government intrusion on individual liberty and decision-making (political and legal frames). This debate also played out in the opinion section of local newspapers. Compared to news stories, there were also more ardent discussions regarding disability rights. Writers, for example, opined that the Schiavo case illuminated just how bad protections for the disabled were in the United States and, occasionally, argued that some segment of the citizenry saw death as preferable to severe disability. This buttresses our argument that the opinion pages are important to democratic deliberate processes insofar as they make discourse more inclusive.

Interestingly, almost all of the significant interactions among the variables involve news format and the tone with which the frame is discussed more specifically (results not shown, available upon request). There are two worth mentioning. First, independently owned outlets are more likely to include editorials featuring passionate discourse on the right to life and medical frames. Compared to news stories in corporately owned outlets, editorials in independently owned outlets were significantly more likely to mention the right to life frame, as opposed to mentioning the right to life frame using a neutral tone (p < .10). Likewise, editorials in independently owned outlets are over three times more likely than news stories in corporately owned outlets to mention the medical frame using a partisan tone, relative to a neutral tone (p < .10). Second, a conservative political ideology also affects the tone with which some of the frames in editorial content are discussed. Compared to articles in conservative states, editorials in both moderate and liberal states are more likely to mention the disability and medical frames using a partisan tone (relative to mentioning each frame using a neutral tone, p ≤ .05 and p ≤ .10 respectively). In other words, we find additional support for the argument that the opinion section of the newspaper provides an important discursive corrective insofar as it provides a space for citizens to argue for and against more particularized ways of understanding the case (Jacobs and Townsley 2011), but that the inclusivity of this forum varies according to an outlet’s ownership and the political ideology of the state. Some opinion pages, in short, are more inclusive than others.

In sum, the multinomial logistic regression analyses offer a more nuanced picture of how ownership, political ideology, geographic region, and news format shape the presence and tone of the frames. We find that ownership affects the presence of, but not the tone with which, frames are included in coverage. In contrast, the political ideology of state significantly shapes the tone, but not frequency, with which frames are discussed in coverage. The geographic region in which a local outlet is located as well as the news format, however, significantly affects both the presence and tone with which diverse ideas are discussed. Taken together, these findings suggest that while local outlets are generally inclusive of diverse viewpoints, whether (and how) a given outlet covers a given idea varies according to the target audience. We discuss the implications of this research to deliberative democratic processes in the concluding section.

DISCUSSION AND CONCLUSION

We find that some sociologists have been overly pessimistic in their assessments of mainstream media coverage. Mainstream news is not the bane of deliberative democratic processes. On the contrary, consistent with current research (Benson 2009, 2013), we find that mainstream newspapers are multiperspectival, meaning they do include diverse ideas and perspectives—regardless of whether the newspaper is independently or corporately owned, regardless of the political leanings of the target audience, and regardless of the geographic location of the outlet. This pessimism, we argue, is the result of scholars emphasizing isomorphism rather than potential diversity in the field. Focusing attention on the hierarchical structure of the field draws attention to how pressures such as ownership and target audience characteristics differentially influence content.

We also find that the opinion pages of the newspaper play an important role in making discourse more inclusive. Opinion pieces and Letters to the Editor not only ensure that a broad range of ideas are included in coverage, but also that the ideas being discussed are done so critically by individuals both supporting and opposing these points of view (Jacobs and Townsley 2011). That said, scholarly concerns regarding the effects of economic factors (ownership and concerns over alienating the target audience specifically) on content are warranted. While local outlets across the country do indeed include diverse points of view in coverage, we find evidence that local outlets shy away from ideas that have the most potential to offend their target audiences. This phenomenon, we find, is not restricted to “hard” news. The most visible effects of audience tailoring, in fact, are on the opinion pages. Clearly, this requires additional research. If tailoring ideas on the opinion pages is widespread, this may damper deliberative processes. The point of the opinion pages, after all, is to expose the citizenry to diverse ideas—some of which should challenge how they think about issues and democracy more generally.

This research makes three contributions to understanding the role of media in deliberative democratic processes. First, it broadens the range of outlets considered relevant to analyses. Scholars tend to define media narrowly (typically as national, mainstream or agenda setting media) and, consequently, draw conclusions about the ability of mainstream outlets to contribute to deliberative processes based on the analysis of a few outlets alone. While we are not arguing that mainstream media provide an ideal public sphere, we are suggesting that local news outlets minimally need to be included in the scholarly conversation. This is particularly true since local outlets remain relevant to citizens even in the digital era and, more importantly, local outlets vary in terms of how inclusive their discourse is. Second, and related, it broadens the range of discourse considered relevant to analyses of deliberative democratic processes. Scholars define discourse narrowly and as a result often miss an important discursive forum—the opinion pages (except see Benson 2013). While it is beyond the scope of this paper to assess whether discourse is civil, this study echoes current scholarship that points to the importance of empirically examining other kinds of mediated content when considering deliberative processes (Gamson and Stuart 1992; Jacobs and Townsley 2011). Clearly, opinion formats provide a space for the inclusion of passionate debate over diverse ideas and deserve additional consideration.

Finally, we take a first step at outlining how economics (ownership and target audience) affect the media content of outlets situated in a hierarchically structured field. Here, we show how Bourdieu’s variant of field theory sheds light on the structure of the field, which simultaneously makes mainstream discourse more inclusive while bending content to avoid offending target demographics at the same time. In short, our findings suggest both that it is too soon to argue that deliberative democratic processes are in peril and that is too soon to discount the effects of economics on newspaper content. It is possible that these findings will be less pronounced in local coverage of other debates, particularly if actors cluster around two opposing sides of an issue. The abortion debate is a good example in this regard. Although there are differences among pro-life and pro-choice groups in terms of their specific goals, some support legal abortion and some do not. Thus, the debate presented in news coverage is relatively clear cut with both sides getting similar rates (and kinds) of coverage (Rohlinger 2007). As a result, it may be more difficult to identify whether there are differences in—and whether some positions are excluded from—the coverage of some debates. While this empirical question deserves additional research, it is clear that social scientists must adopt a more nuanced approach to analyzing discourse and the forces that shape it.

Biographies

Deana A. Rohlinger is an associate professor in the Department of Sociology at Florida State University. Her work examines mass media, social movements, and political participation. She is author of Abortion Politics, Mass Media, and Social Movements in America (Cambridge University Press) as well as pieces published in Social Problems, Sociological Theory, Mobilization, The Sociological Quarterly, Research in Social Movements, Conflicts and Change, The American Behavioral Scientist, Social Movement Studies, and several edited volumes. Her current research includes an examination of the relationship between Internet usage and political participation in conservative and progressive movements as well as the role of organizational reputation in social movement outcomes.

JoEllen Pederson is an Assistant Professor of Sociology at Longwood University. Her research focuses on welfare states in a cross-national context, aging, and public policy.

Giuseppina Valle is a postdoctoral fellow at the University of Texas at Austin in the Population Research Center. She received her PhD in sociology from the Florida State University in 2013. Her primary research interests include social demography, adolescent and young adult well-being, sexual behavior, and quantitative methodology.