Representation of Advocacy Groups' Ideology in Newspapers

JMQXXX10.1177/1077699012455385Journ 455385alism & Mass Communication QuarterlyMcCluskey and Kim

Journalism & Mass Communication Quarterly XX(X) 1–20 Moderatism or © 2012 AEJMC Reprints and permission: Polarization? sagepub.com/journalsPermissions.nav DOI: 10.1177/1077699012455385 Representation of http://jmcq.sagepub.com Advocacy Groups’ Ideology in Newspapers

Michael McCluskey1 and Young Mie Kim2

Abstract Scholars and commentators argue that the United States has become politically polarized in recent years, with news content itself favoring polarized views. If true, this represents a radical shift from Gans’s enduring news value of moderatism. By examining 208 advocacy groups’ ideology and their representation in 118 newspapers, this study revisits Gans’s moderatism argument and investigates polarization in news content. Analysis demonstrates that moderate groups had less prominence within articles, with no differences in tone. Polarization may offer a higher news value by presenting inherent conflict and a means for journalistic balance.

Keywords Moderatism, polarization, news, advocacy groups

Since Herbert Gans’s sociological investigation of news organizations was published in 1979,1 scholars have believed that certain journalistic values influence news decisions. Gans identified eight “enduring values” that subconsciously guide news decisions—among them moderatism,2 which discourages excess or extremism in human activities, including politics. Interestingly, even Gans suggested that the “enduring values are not timeless, and they may change somewhat over the years.”3 This suggests the “enduring” value of

1Ohio State University, Columbus, OH, USA 2University of Wisconsin–Madison, Madison, WI, USA

Corresponding Author: Michael McCluskey, School of Communication, Ohio State University 3072 Derby Hall, 154 North Oval Mall, Columbus, OH 43210-1339, USA Email: [email protected] 2 Journalism & Mass Communication Quarterly XX(X) moderatism should be systematically reassessed in this changing media environment. Since Gans’s book was published in 1979, our media environment has fundamentally transformed, with technological innovations like the Internet expanding the breadth of information sources, shrinking the market of traditional mass-audience-oriented news outlets. Bennett and Iyengar propose we have entered an era of information “stratamentation”—a simultaneous combination of media stratification and fragmentation— that reflects audience media selection and news organization content or ideological targeting of audience segments.4 Indeed, contemporary social scientists claim that as the U.S. social and political environment has become more polarized, news media prefer polarized over moderate views. Media observers offer anecdotal evidence of polarization in news content,5 while scholars argue not only that polarization in news has taken root, but also that it has affected public attitudes toward the political and social world.6 Polarization may have a spiraling effect in which elite polarization, amplified by politically divided news and commentary, creates even larger divisions in the mass public that feed into support for increasingly polarized political elites. Although much attention has been paid to the balance of conservative and liberal views in the news,7 scholars have paid little attention to comparing polarized and moderate views. To fill this void, this study raises a fundamental question regarding news values in the contemporary media environment: Does newspaper content favor politically polarized views over moderate views? To address this question, we examined news coverage mentioning any of 208 issue groups, including public advocacy groups (e.g., ACLU, NRA), professional/trade associations, and labor unions that represent a range of political ideologies. Groups were randomly selected from Internal Revenue Service (IRS) databases. With this unique data set, we investigated differences in news presence and portrayal between groups advocating moderate political views and those advocating polarized views. This study introduces multidimensional measures of presence to capture not only which actors appear in the news, but also audience perspec- tives, such as how many people are likely to be exposed to a news story and the potential likelihood that the audience will notice those actors within a news story.

Dominant Journalism Practices: Moderatism In explaining news production, Gans proposed eight “enduring values” reflecting systematic patterns in news content, granting more attention to certain positions, issues, and actors. Gans explained that enduring values “are values that can be found in many different types of news stories over a long period of time; often, they affect what events become news, for some are part and parcel of the definition of news.”8 Furthermore, Gans explained that the enduring “values in the news are rarely explicit and must be found between the lines—in what actors and activities are reported or ignored, and in how they are described.”9 Moderate solutions are favored over those in the extremes, and “their values rarely coincide with those on the Far Right or the McCluskey and Kim 3

Far Left.”10 Similarly, Shoemaker and Reese asserted, “Fanaticism of any sort is treated as suspect, as is conspicuous consumption and fervent political ideology.”11 It is important to recognize that enduring news values are embedded within journalistic traditions emphasizing objectivity, fairness, and balance within stories. The enduring news values are rarely explicit and can best be seen by looking retrospec- tively at which positions, actors, and issues populate the news. That is, characteristics of who gets portrayed in the news provide clues into the enduring values of news organization and also set the information context for the public agenda. Therefore, it is critical to look at moderatism less as a guiding principle that journalists use to determine what is newsworthy and more as a way to explain common and systematic patterns of coverage. Moderatism influences the media industry as a mechanism of social control, in which media identify threats to the status quo and portray deviant ideas in ways calculated to underscore their deviance.12 Instead of the traditional role as a watchdog, media may operate as a guard dog13—maintaining order and protecting society against potential internal and external threats. Moderatism suggests that advocacy groups pursuing moderate views receive greater presence in news coverage than advocates for polarized views. One concern of advocacy groups pursuing less moderate goals is that journalists may see them as excessive, extremist, and dangerous. That is, journalists, as mechanisms of social control, would be skeptical of threats to the establishment and act to defend the social order. Ironically, little research has established moderatism as an enduring news value. We found fifty-six separate works that cited Gans and/or Deciding What’s News and moderatism, including journal articles, books, conference papers, and theses.14 All available works written in English were reviewed, and most listed the enduring values with little or no elaboration.15 Several others provided descriptive evidence of moderatism.16 Just three explicitly compared moderatism with polarized views, one evaluating a newspaper and the others, television. A conference paper used independent measures of ideology in finding that U.S. senators holding moderate views received more coverage in the Washington Post than those holding extreme positions.17 Schiffer noted, “Apparently the norm of moderation helps journalists set the ideological boundaries, outside of which a senator ventures at the expense of issue association.”18 A study of the television news commentary program Nightline found that guests and topics primarily represented centrist political interests, creating a false impression of the range of the domestic political scene.19 By contrast, Kuklinski and Sigelman found that rela- tively extremist senators received more network television coverage than more moderate peers.20 They explained, “Because this contrapunctual format works best when positions are contrasted in their purest form, it should enhance the journalistic appeal of senators who espouse extreme positions, i.e., ideologically extreme senators and those extreme in their support for or opposition to the president.”21 In short, evidence of Gans’s enduring value of moderatism is limited to descriptive examples and studies with contradictory findings. 4 Journalism & Mass Communication Quarterly XX(X)

Changing Political and News Environment

Three patterns in the current news landscape support revisiting and reevaluating moderatism in contemporary journalism practices: political polarization, the rise of advocacy groups and group polarization, and media polarization. Political Polarization. Political observers, scholars, and news commentators contend that the United States has shifted toward political polarization among political elites,22 the mass public,23 and political parties and activists.24 Polarization is typified by divisions like red state/blue state,25 a perceived inability for Democrats and Republicans to work together,26 and evidence of selective exposure.27 Public debate over a health care bill in Congress in 2010 was one example of polarization. Discourse seemed dominated by arguments that the legislation would offer meaningful change and financial stability to the health care system, or it would lead the country into financial ruin and take away freedom of choice. Party unity was the “overriding imperative,”28 and compromise was missing, replaced with what Baker calls “an ugly zero-sum game mentality.”29 Scholars agree elite polarization has existed for two decades,30 while the evidence for mass polarization is mixed. Parties in the electorate have become more distinct,31 especially on racial,32 moral,33 and cultural issues34—the types of issues that often engage advocacy groups. Ideological conflict has increased among political elites, moving Democrats left and Republicans right.35 The contemporary political system encourages activists to take more extreme views on multiple issue dimensions, facili- tating “conflict extension”—polarization of political parties.36 Ideological polarization has increased among the public as well,37 and the percentage of the public at the ideological center shrank from 41% to 28% since the mid-1980s.38 However, others dis- pute the mass polarization trend, finding a general pattern of the public as centrist,39 moderate, or ambivalent in ideological orientation. Rise of Advocacy Groups and Group Polarization. Terms such as interest groups, activist groups, pressure groups, and nongovernmental organizations are used to label the phenomenon of individuals acting collectively for social or political causes, often in formal groups with paid staff, large budgets, and bureaucratic structures. Advocacy groups thus occupy a critical middle ground between political elites and the mass public, representing the views of a particularly engaged segment of the public. Political polarization theorists, both supporters and critics of the mass polarization hypothesis, agree that the deepest divisions are among the most informed and active citizens,40 and that those who care deeply tend to hold extreme views.41 Passionate, ideologically polarized, and active individuals increasingly work together. Since 1960, the United States has seen an explosion in the number of advocacy groups,42 placing more impor- tance on understanding their interactions with news organizations. Advocacy groups, by pursuing a narrow range of political and social goals, may push polarization.43 A small group of activists on the left and right may exert influence out of proportion to their numbers, and it is the elites and activists who are polarized.44 McCluskey and Kim 5

The myth of a culture war rests on “self-serving misrepresentation by issue activists, and selective coverage by an uncritical media more concerned with news value than with getting the story right.”45 News, Polarization, and Advocacy Groups. According to Bennett and Iyengar, media in the United States are polarized, this polarization has accelerated in recent years, and this trend has changed the foundations of political communication.46 This trend affects media usage patterns, and those trends affect perceptions about the political world, represented by increasing polarization in public opinion. Ben- nett and Iyengar suggested that a more diversified media environment offers the public more opportunities to choose media content matching their preexisting beliefs47—e.g., selective exposure48—or, alternately, that these choices reflect the hostile media phenomenon as the audience avoids material contrary to their beliefs.49 A nationwide survey found those with a more polarized political ideology were more likely to prefer news from their own point of view.50 Furthermore, news organizations have a strong economic incentive to cater to the audience’s political preferences.51 Media content, by presenting polarized views, may give the audience a false sense of the political world. One media commentator suggests, “To view and to listen is to become convinced that there are only two, diametrically opposed philosophical approaches to the issues. . . . Opinions from the middle are underrepresented, even shunned, in the modern debate. . . . The more doctrinaire the viewpoint, the better the odds it will be heard.”52 Despite the indispensable implications of group politics in the changing media landscape, only in recent years have scholars systematically examined the interplay among advocacy groups, their strategic activities, and news coverage.53 For advocacy groups, news content is a valuable communication resource as a means to influence public opinion and public officials. While influencing public officials may be the most direct path to political change, not all activists have access,54 so news attention becomes an even more critical resource. News coverage raises awareness among public officials and citizens about advocacy organizations, problems the organizations attempt to address, and legitimacy of organizations to address those problems. Advocacy groups receiving news attention potentially can mobilize public support and exert pressure for political change.55 Legitimacy that goes with news attention may further give the organization access to closed doors of public officials and subsequent news attention. In essence, the struggle for media attention becomes an element in the battle for political influence and control.56 Advocacy groups face two challenges—first, gaining attention from journalists and, second, news coverage that accurately portrays their goals. The focus of this study is on (1) coverage, variously called presence, voice, or the “status conferral function”57 of mass media, and (2) tone, reflected in the way organizations are portrayed within the balance of each article. The absence of coverage—described as “symbolic annihilation”58 or nonevents for activities that received no news 6 Journalism & Mass Communication Quarterly XX(X) coverage59—suggests that the organization and its goals may never reach the public agenda. In addition, tone potentially increases the possibility of change in social practices or in public policies,60 and reflects how issues are defined and interpreted.61 As the first step in exploring moderatism and polarization in news, our focus is on the presence and tone in newspaper content of advocacy groups espousing polarized political ideology with those holding moderate views. Presence is a crucial indicator of news selection, as it is linked to processes such as gatekeeping and coverage practices.62 News presence provides enormous normative implications in giving the audience a perception of “what is out there” (or existence), while tone suggests how the audience may perceive groups. As we were able to identify group ideologies, we can observe differences within types of advocacy groups, as measured by both presence and tone. Presence was examined through multiple dimensions by looking at (1) the amount of coverage each type of organization received; (2) the size of the potential audience, as reflected in newspaper circulation; (3) the prominence of stories, as measured by where articles appeared in the newspaper; (4) the prominence of the advocacy group within stories, measured by where and how often the organization appeared in each story; and (5) variations by newspaper size, the scope of the organization, and the issues pursued by the organization. Tone was measured by evaluating how each group was portrayed within each article, evaluated by the overall balance of the article.

Method First, 1,176 nonprofit public advocacy groups, labor unions, professional associations, and trade associations were randomly drawn from IRS databases. Next, the executives who oversee group activities and set missions and strategies were interviewed by using the Computer Assisted Telephone Interview system. Ultimately, 242 interviews were obtained between October 13 and December 12, 2006, yielding a response rate of 35.9%.63 The final analysis used 208 cases, excluding 31 partially completed interviews, 2 interviews conducted incorrectly, and 1 organization that could not be searched for newspaper content, because its name consisted of three common words that the respective databases were unable to search as a word string. Entering the name as a word string in LexisNexis returned more than three thousand matches for a single day.

Characteristics of the Population Group ideology. Based on self-reports, each organization assessed its ideological position within its issue of interest, on a 7-point scale ranging from 1 = very conservative to 7 = very liberal.64 Overall, the 208 groups evaluated their ideology as more liberal than conservative. The 32 conservative organizations included 20 very conservative, 2 conservative, and 10 leaning conservative. Another 105 assessed themselves as liberal, including 41 very liberal, 23 liberal, and 41 leaning liberal. The remaining 71 groups placed themselves in the moderate category. McCluskey and Kim 7

Newspaper content. The population of 4,304 newspaper articles was discovered in several steps. The 118 newspapers were selected based on newspaper size (the 20 largest general-interest daily newspapers in the United States by weekday circulation65) and proximity to the advocacy organization headquarters (the largest daily newspaper within fifty miles of each group, or, for groups without a daily newspaper within fifty miles, the nearest daily newspaper). The sampling design increased the chance organizations pursuing locally oriented goals would be captured in the news content, as the nearest daily newspaper would be more likely than nonlocal newspapers to mention the organization. The 118 newspapers were geographically diverse (forty-two states) and ranged in weekday circulation from 2.2 million to 2,849. The name of each organization with a completed survey was entered in a full-text search of article databases for articles published in 2006, with forty-three newspapers searched using LexisNexis and the remainder in the archives of individual newspapers. Database searches turned up 5,576 potential articles. Articles in which the group name did not appear, either because it matched a common phrase or the database would not allow a search of a phrase, were excluded, as were duplicate articles, brief articles (fewer than five sen- tences), and death notices. That left a population of 4,304 articles.

Characteristics of the Sample Group ideology. For analysis, the twenty very conservative and forty-one very liberal groups were identified as polarized and were compared with the seventy-one moderate groups.66 Although polarization scholars typically conceptualize a bimodal variable of two extremes,67 the focus of this research compares political extremes to the middle.68 Scope of group activity. Group representatives were asked if their activities were primarily at the international, national, or state/local level. Among the polarized groups, fifteen were international, fourteen were national, and thirty-two were state/ local. Among the moderate groups, fourteen were international, twenty-six were national, and thirty-one were state/local. Main issue of interest. Representatives were asked whether or not their organization addressed each of six issues of interest. Among the polarized groups, fourteen addressed abortion, two economic issues, ten GLBT (gay, lesbian, bisexual, transgender) issues, twelve civil rights, three education, zero environmental, four labor, and four women’s issues, with the remaining twelve dealing with other issues. Among the moderate groups, two addressed abortion, one economic issues, three GLBT issues, twenty-two civil rights, seven education, two environmental, two labor, and four women’s issues, with the remaining twenty-eight dealing with other issues. Newspaper content. A sample (N = 1,077) was drawn in two steps for analysis, with a goal to draw representation from each specific organization. First, up to five articles per organization were randomly drawn, meaning that each organization mentioned in news coverage was represented in the sample. Next, up to half of the next ten articles for each organization were randomly selected. Thus, no more than ten articles were coded from any specific organization. 8 Journalism & Mass Communication Quarterly XX(X)

Two trained coders analyzed the news content. Intercoder reliability testing was done on ninety-seven articles using Krippendorff’s alpha, which takes chance agreement into consideration. The filter variable—assessing whether the group appeared in the article—was reliable (α = .87), as were the name of the organization (α = 1.00), the name of the newspaper (α = .97), and whether the article was non-opinion-based or opinion-based (α = .82). Several variables were used to measure dimensions of presence, including location in the newspaper (coded as 1 = first page, first section, 2 = first page, other section, 3 = inside page, first section, 4 = inside page, other section, and 0 = unknown; α = .87), the number of paragraphs in the article (α = .96), whether the group was mentioned in the first paragraph (α = .89), the paragraph in which the group was first mentioned (α = .90), and the number of paragraphs in the article in which the group appeared (α = .80). Two variables were calculated: depth of the first mention of the group in the article (ratio of the paragraph of first mention divided by the number of paragraphs in the article) and percentage of paragraphs in which the group is mentioned (ratio of the number of paragraphs in which the group was mentioned divided by the number of paragraphs in the article). Two other variables used in measuring presence were weekday circulation of the newspaper, drawn from Editor & Publisher,69 and the total number of articles in which each group was mentioned. Portrayal of group was measured on a 5-point scale (1 = very negative, 2 = somewhat negative, 3 = neutral, 4 = somewhat positive, 5 = very positive), assessing how the group was portrayed within the overall balance of the article. The “very” categories applied only when the overall thrust of the article specifically praised or denigrated the advocacy group. The measure was reliable (α = .91).

Results Descriptive Data

Population. The population consisted of 4,304 articles (M = 20.7 articles per group, Mdn = 5), with one organization appearing in 1,332 articles and forty-seven appearing in no articles. Another sixty groups were in 1 to 5 articles. The majority of articles were nonopinion (90.7%), and the mean weekday circulation of the newspaper was 374,804. As for location in the newspaper, 13.5% were on the front page of the first section, 20.4% were on another section front, 22.5% were on an inside page of the first section, 36.2% were on an inside page in other than the first section, and the remaining 7.6% were not listed. The newspapers with the most articles were the Los Angeles Times (n = 202) and the Washington Post (n = 192). Although the sixty-one polarized organizations received more coverage (n = 1,942 articles, M = 31.84 articles per organization, SD = 169.84) than moderates (71 groups, n = 1,101 articles, M = 15.51, SD = 20.80), the difference was not significant, t(130) = 0.80, ns, indicating the amount of news coverage did not statistically differ between the polarized and moderate groups. McCluskey and Kim 9

Table 1. Mean Characteristics of Coverage by Organizational Ideology Polarized groups Moderate groups M SD M SD F Circulation (weekday) 297,900 307,724 384,985 399,202 8.31** Depth of first mention of group by 0.48 0.30 0.54 0.31 5.92* % of paragraphs % of paragraphs in which group 0.20 0.23 0.14 0.19 11.93*** mentioned Tone: Portrayal of group 3.25 0.56 3.32 0.54 1.87 N = 567 264 303 Location in newspaper 2.97 1.09 2.77 1.19 3.53 N = 496 206 290

Polarized groups consist of those self-identified as very conservative or very liberal. Moderate groups consist of those self-identified as neither conservative nor liberal. Portrayal of group: 1 = very negative, 2 = negative, 3 = neutral, 4 = positive, 5 = very positive. Location in newspaper excludes articles in which location is unknown. Coded as 1 = first page, first section, 2 = first page, other section, 3 = inside page, first section, 4 = inside page, other section. *p < .05. **p < .01. ***p < .001.

Presence in News Coverage

The following analyses are confined to the 567 nonopinion articles (analyses including opinion articles resulted in findings that varied little from the findings presented below) mentioning either the polarized or moderate groups. Overall, moderates received less presence in news coverage than polarized groups (see Table 1), providing support for the polarization hypothesis. ANOVA testing showed moderates were mentioned first later in the articles and received less coverage within the articles, although moderates did appear in larger newspapers. ANOVA tests showed no differences in location within the newspaper, and chi-square analysis showed differences just short of statistical significance on frequency of the group in the first paragraph, with polarized groups in the lead paragraph in 14.4% of articles compared to 10.5% for moderates (χ2[1, N = 570] = 2.04, p = .09). Overall, the results demonstrate more support for the polarization hypothesis than for moderatism.

Tone in News Coverage Although polarized groups received more prominent presence in news articles, this characteristic could be undermined if the groups were portrayed in a negative way. Evaluation of tone becomes an important checkpoint to determine whether presence is meaningful. Analyses showed that the polarized and moderate groups did not significantly differ in tone of portrayal (see Table 1), as both were somewhat positive. Thus, at the aggregate level, polarized groups received much greater presence in the news than moderates, but no differences were found in tone. 10 Journalism & Mass Communication Quarterly XX(X)

Newspaper Size, Group Scope, and Group Issues

Additional analyses examining characteristics by newspaper size, by the scope of the group’s focus, and by characteristics within specific issues pursued by groups offered additional evidence of polarization. A median split of newspapers by circulation showed polarized groups were mentioned in a higher mean level of paragraphs than moderates in both large and small newspapers (see Table 2), and polarized groups were mentioned significantly earlier in stories, based on mean depth of first mention in small newspapers. Additional differences were found among advocacy groups with a state/local scope, in which polarized groups were mentioned first earlier in stories and were mentioned in a higher mean level of paragraphs than moderates (see Table 3). In addition, polarized groups (17.2%) appeared more often in the first paragraph (χ2[1, N = 308] = 2.37, p = .08) than moderates (11.0%). On the main issue pursued by the groups, just two of the eight issues had enough of an ideological distribution to analyze. Among groups engaged in civil rights issues, polarized groups were mentioned earlier in stories (M = 0.43 vs. 0.54, F[1, 168] = 4.96, p = .03) and in a higher mean percentage of paragraphs (M = 0.22 vs. 0.14, F[1, 168] = 6.75, p = .01) than moderates. No differences were found among groups engaged in abortion issues. In contrast, moderates with a state or local scope appeared in larger newspapers than polarized groups, and moderates had significantly more positive tone than polarized groups in large newspapers and when they had a national scope.

Discussion Analysis shows more evidence of polarization than moderatism in the representation of advocacy groups’ ideology in news, suggesting that the “enduring value” of moderatism identified by Gans in 1979 may no longer apply in the contemporary media and political environment. In analyses of presence and tone, polarized groups had favorable treatment more often than moderates, supplying evidence for the polarization hypothesis. Moderate groups had less prominence within articles, while no differences were found in location within the newspaper or in tone of portrayal. Additional evidence for polarization was found by newspaper size, by group scope, and within specific issues that the groups pursue. Furthermore, differences were not due to group characteristics, as group-level analyses indicated that polarized and moderate groups did not significantly differ in number of members, expenses, revenue, staff size, presence of public relations staff, or proximity to Washington, D.C. Moderatism as an enduring news value has received little systematic analysis in the past three decades, with previous evidence for moderatism drawn from a study of one newspaper,70 and from one television program,71 and support for polarization from network television.72 The growth and diversity of the news and information environment since 1979, coupled with Gans’s admission that the enduring values may not be timeless,73 made reevaluation of moderatism timely, especially given the fact that scholars have argued that news coverage in recent years has become more polarized.74 McCluskey and Kim 11

Table 2. Mean Characteristics of Coverage by Organizational Ideology and Newspaper Size Polarized Moderate groups groups M SD M SD F Depth of first Large 0.50 0.29 0.53 0.31 0.53 mention of newspapers group by % of paragraphs Small 0.47 0.30 0.56 0.31 7.08** newspapers

% of paragraphs Large 0.18 0.22 0.13 0.18 3.87* in which group newspapers mentioned Small 0.22 0.23 0.15 0.19 7.29** newspapers

Portrayal of group Large 3.14 0.53 3.30 0.52 5.81* newspapers

Small 3.34 0.56 3.33 0.56 0.01 newspapers

Large newspapers: N = 260 113 147 Small newspapers: N = 307 151 156 Location in Large 3.02 1.00 2.78 1.16 2.87 newspaper newspapers

Small 2.91 1.19 2.76 1.22 0.88 newspapers

Large newspapers: N = 249 111 138 Small newspapers: N = 247 95 152

Polarized groups consist of those self-identified as very conservative or very liberal. Moderate groups consist of those self-identified as neither conservative nor liberal. Newspaper size determined by median split, with small newspapers less than 300,000 weekday circulation. Portrayal of group: 1 = very negative, 2 = negative, 3 = neutral, 4 = positive, 5 = very positive. Location in newspaper excludes articles in which location is unknown. Coded as 1 = first page, first section, 2 = first page, other section, 3 = inside page, first section, 4 = inside page, other section. *p < .05. **p < .01.

Considering the breadth and depth of the data—survey data from 208 advocacy groups, combined with data from a sample of newspaper articles drawn from 118 newspapers—the analysis fairly represents coverage of ideologically driven advocacy organizations in newspaper content. 12 Journalism & Mass Communication Quarterly XX(X)

Table 3. Mean Characteristics of Coverage by Organizational Ideology and Scope of Group Activities Polarized groups Moderate groups M SD M SD F Circulation International 529,745 531,825 533,960 607,630 0.00 National 329,031 191,847 415,902 435,454 2.20 State/local 229,356 231,615 315,646 252,055 9.90**

Depth of first International 0.46 0.28 0.57 0.29 2.72 mention of National 0.53 0.30 0.52 0.30 1.00 group by % of State/local 0.47 0.30 0.56 0.33 6.77** paragraphs

% of paragraphs International 0.18 0.20 0.14 0.17 0.74 in which group National 0.18 0.19 0.14 0.19 1.63 mentioned State/local 0.22 0.25 0.15 0.19 7.70**

Portrayal of International 3.30 0.52 3.40 0.58 0.62 group National 3.11 0.55 3.32 0.54 5.68* State/local 3.29 0.57 3.29 0.53 0.00

International scope: N = 83 40 43 National scope: N = 177 61 116 State/local scope: N = 307 163 144

Location in International 2.92 0.97 2.83 1.15 0.14 newspaper National 2.98 1.12 2.74 1.23 1.62 State/local 2.97 1.13 2.78 1.18 1.78

International scope: N = 80 38 42 National scope: N = 168 58 110 State/local scope: N = 248 110 138

Polarized groups consist of those self-identified as very conservative or very liberal. Moderate groups consist of those self-identified as neither conservative nor liberal. Scope self-identified by group representatives. Portrayal of group: 1 = very negative, 2 = negative, 3 = neutral, 4 = positive, 5 = very positive. Location in newspaper excludes articles in which location is unknown. Coded as 1 = first page, first section, 2 = first page, other section, 3 = inside page, first section, 4 = inside page, other section. *p < .05. **p < .01.

Contemporary news values offer reasons for news presence favoring polarization. Those advocating polarized views offer a higher potential for conflict, a common news value. For instance, “Disagreement, division, polarization, battles, and war make better copy than agreement, consensus, moderation, cooperation, and peace. Thus, the culture war frame fits the news values of journalists who cover American politics.”75 Smerconish provides vivid anecdotes of the thirst for polarized views while meeting McCluskey and Kim 13

separately with producers for CNN’s Larry King Live and Fox News, who each wanted him to make one-sided conservative comments prior to the 2008 presidential election and were dismayed by his mixed views. As he explained, “The message of both episodes is clear: There is no room for nuance. Either you offer a consistent (possibly artificial) ideological view or you often don’t get a say.”76 Second, extremes are more intuitively novel, entertaining, and colorful, representing another common news value. This media trend is best represented by the use of vivid exemplars77—think of colorful characters or highlighting people in costumes— to fit a prior assumption of a polarized country, creating anecdotes that stick in the minds of the audience. Moderate voices may be more difficult to portray as exciting than extreme voices. Third, news organizations traditionally espouse normative principles of objectivity, fairness, and balance. Polarized views provide more clearly identified “two sides of a story” that represent balance. Cook made a similar point, suggesting that reporters like stories “if they can be dramatized by two (and only two) distinct sides and an easily reported conflict.”78 Selecting two opposing views—say very conservative and very liberal—thus becomes a shortcut for a range of views. In this way, journalists appear to have a “third-person” view maintaining objectivity as a professional value.79 Moderate views, on the other hand, simply are less easy to define as representing a viewpoint than polarized left–right opinions. The large number of articles analyzed points toward an overall trend of balance among ideological extremes. Evidence for polarization was clear by presence in the news, but not for tone, in which polarized and moderate groups were treated similarly. Additional evaluation of article tone revealed interesting patterns: While no differences were found comparing polarized and moderate groups, moderates and liberals received significantly more positive portrayals than conservatives. However, again, it should be emphasized that the overall balance was somewhat positive for all ideological types. Further investiga- tions revealed that those differences were domain-specific: While conservative groups pursuing abortion goals were somewhat positive (M = 3.09, SD = 0.57), the sheer number of articles (n = 79) dragged down the overall tone. The moderates engaged in abortion (n = 12 articles, M = 3.00, SD = 0.43) were even less positive than the conservatives, although the difference was not statistically significant, and no liberal groups in the data set were engaged in addressing abortion. Analyses eliminating groups engaged in abortion showed no significant differences (actually, the conservatives, at 3.38, had a slightly higher mean score than either moderates or liberals). This neutral tone in abortion stories makes sense, as the issue has long pitted two distinct and engaged sides, so journalistic balance within stories is not only normative, but perhaps pragmatic to avoid criticism. Besides contributing evidence of a polarized news environment, another major contribution of this study is its multiple dimensions of presence, which expand and add depth to a measure that has often been operationalized as either present or absent. Presence has long been recognized by scholars as an important indicator of potential public perception, as news coverage confers status,80 portraying those actors as important and legitimate, potentially fueling the group’s success in pursuing its social and 14 Journalism & Mass Communication Quarterly XX(X) political agenda.81 By contrast, the absence of coverage keeps advocacy groups almost entirely out of the public eye.82 The more precise measures of presence introduced here consider both the size of the news audience and what readers may experience in reading stories. Polarized groups appeared earlier in articles and in more paragraphs, suggesting more people had the opportunity to note those groups, fueling perceptions of those groups as important or legitimate.83 One limitation of this study was its focus on newspapers, as television remains a primary source of news for many individuals. Considering that much of the anecdotal evidence discussed polarization in television news, even starker evidence might have been found by analyzing television. In addition, the findings presented here can be enhanced by providing more depth about how the advocacy groups were portrayed in the articles. This would inform advocacy groups in designing media strategies, relations between journalists and advocacy groups, news norms in journalism as related to advocacy groups and political polarization, and the news environment experienced by the public. Although the presence of advocacy groups in news content is not the only way to measure polarization, it does offer some useful insights to a broader trend that may profoundly affect public perceptions about the political world.

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. Herbert J. Gans, Deciding What’s News: A Study of CBS Evening News, NBC Nightly News, Newsweek and Time (New York: Pantheon Books, 1979). 2. The other seven enduring values are ethnocentrism, altruistic democracy, responsible capi- talism, small-town pastoralism, individualism, the preservation of social order, and the need for national leadership. 3. Gans, Deciding What’s News, 41. 4. W. Lance Bennett and Shanto Iyengar, “A New Era of Minimal Effects? The Chang- ing Foundations of Political Communication,” Journal of Communication 58 (4, 2008): 707–31. Bennett and Iyengar provide evidence for a historic shift in media fragmentation, increased supply of information, and strategic choices by news organizations to target niche audiences. 5. E. J. Dionne, Jr., “One Nation Deeply Divided,” Washington Post, November 7, 2003, A31; Thomas Friedman, “Two Nations under God,” New York Times, November 4, 2004, A25; Michael Smerconish, “The Media’s Black-and-White World,” Washington Post, June 11, 2010, A17. 6. Bennett and Iyengar, “New Era of Minimal Effects?” McCluskey and Kim 15

7. Matthew Gentzkow and Jeremy M. Shapiro, “Media Bias and Reputation,” Journal of Political Economy 114 (2, 2006): 280–316; Tim Groseclose and Jeffrey Milyo, “A Mea- sure of Media Bias,” Quarterly Journal of Economics 120 (November 2005): 1191–1237. 8. Gans, Deciding What’s News, 41. 9. Gans, Deciding What’s News, 39–40. 10. Gans, Deciding What’s News, 68. 11. Pamela J. Shoemaker and Stephen D. Reese, Mediating the Message: Theories of Influ- ences on Mass Media Content, 2nd ed. (New York: Longman, 1996), 83. 12. Shoemaker and Reese, Mediating the Message. 13. George A. Donohue, Phillip J. Tichenor, and Clarice N. Olien, “A Guard Dog Perspective on the Role of Media,” Journal of Communication 45 (2, 1995): 115–32. 14. A search of Google Scholar turned up 1,292 times that Deciding What’s News has been cited, and Communication & Mass Media Complete found 682 citations. After filtering for the term moderatism, 33 were found via Google Scholar, 19 via Communication & Mass Media Complete, and 4 unique additional citations in JSTOR. 15. Among the journal articles were David L. Altheide, “Moral Panic: From Sociological Con- cept to Public Discourse,” Crime Media Culture 5 (1, 2009): 79–99; Jay Black and Ralph Barney, “Commentary: Journalism Ethics since Janet Cooke,” Newspaper Research Jour- nal 13–14 (fall 1992/winter 1993): 2–16; Edward Caudill, “Intelligently Designed: Cre- ationism’s News Appeal,” Journalism & Mass Communication Quarterly 87 (spring 2010): 84–99; Thomas Bryan Christie, “The Role of Values in Predicting Talk Radio Listening: A Model of Value Equivalence,” Journal of Radio Studies 14 (1, 2007): 20–36; Heloiza Golbspan Herscovitz, “Lula vs. Larry Rohter: Misconceptions in International Coverage,” Brazilian Journalism Research 3 (1, 2007): 155–72; Jack Johnson-Hill, “Unheard Voices: Jamaica’s Struggle and the Multinational Media,” Caribbean Quarterly 27 (June/September 1981): 1–20; Penelope Ploughman, “The American Print News Media ‘Construction’ of Five Natural Disasters,” Disasters 19 (December 1995): 308–26; Ida Schultz, “The Journalistic Gut Feeling: Journalistic Doxa, News Habitus and Orthodox News Values,” Journalism Practice 1 (May 2007): 190–207; Jerome H. Skolnick and Candace McCoy, “Police Accountability and the Media,” American Bar Foundation Research Journal 9 (summer 1984): 521–57; Elizabeth K. Viall, “Measuring Journalistic Values: A Cosmopol- itan Community Continuum,” Journal of Mass Media Ethics 7 (1, 1992): 41–53; Jonathan Weinberg, “Broadcasting and Speech,” California Law Review 81 (October 1993): 1101, 1103–1206; Lee Wilkins, “Taking the Future Seriously,” Journal of Mass Media Ethics 5 (2, 1990): 88–101; and Lee Wilkins, “Between Facts and Values: Print Media Coverage of the Greenhouse Effect, 1987–1990,” Public Understanding of Science 2 (1, 1993): 71–84. 16. Examples include the following: the New York Times was less supportive of “sabre rat- tlers” in Israel and the PLO, in Raymond Stock, “Prestige Press at War: The New York Times and Le Monde in Lebanon, August 1–September 26, 1982,” Middle East Journal 39 (summer 1985): 317–40; media coverage of 1983 confrontations in Waco between Branch Davidians and federal agents in David Altheide, An Ecology of Communication: Cultural Formats of Control (New York: Aldine de Gruyter, 1995), and Philo C. Washburn, The Social Construction of International News: We’re Talking About Them, They’re Talking 16 Journalism & Mass Communication Quarterly XX(X)

about Us (Westport, CT: Praeger, 2002); newspapers’ coverage of death-row exonerations, in David Niven, “Southern Newspaper Coverage of Exonerations from Death Row,” Journal of Criminal Justice and Popular Culture 11 (spring/summer 2004): 20–31; stories written by Catholic and evangelical journalists, in John Schmalzbauer, “Telling Catholic and Evangelical Stories in American Journalism: The Impact of the Religious Imagina- tion,” U.S. Catholic Historian 20 (spring 2002): 25–44; coverage by Spanish writers of democracy and dissent, in Iñaki Garcia Blanco, “Fostering Conflict or Consensus? Nor- mative Conceptions of Democracy in the Spanish Press” (paper, Equal Opportunities and Communication Rights: Representation, Participation, and the European Democratic Deficit symposium, Brussels, October 11–12, 2007); use of bioethicists as expert sources in newspapers, in Marjorie Kruvand, “Bioethicists as Expert Sources in Science/Medical Reporting,” Newspaper Research Journal 30 (summer 2009): 26–41; and television coverage of the 1994 Winter Olympics, in Steven Nordenstedt, “Enduring Values: A Content Analysis of Television Coverage of the 1994 Olympic Games” (master’s thesis, San Jose State University, 1995). 17. Adam Schiffer, “Explaining Policy Coverage of U.S. Senators” (paper, Midwest Political Science Association annual meeting, Chicago, 2002). 18. Schiffer, “Explaining Policy Coverage,” 18. 19. David Croteau and William Hoynes, By Invitation Only: How the Media Limit Political Debate (Monroe, ME: Common Courage, 1994). 20. James H. Kuklinski and Lee Sigelman, “When Objectivity Is Not Objective: Network Television News Coverage of U.S. Senators and the ‘Paradox of Objectivity,’” Journal of Politics 54 (August 1992): 810–33. 21. Kuklinski and Sigelman, “When Objectivity Is Not Objective,” 815. 22. Morris P. Fiorina and Samuel J. Abrams, “Political Polarization in the American Public,” Annual Review of Political Science 11 (2008): 563–88; Shawn Treier and D. Sunshine Hillygus, “The Nature of Political Ideology in the Contemporary Electorate,” Public Opin- ion Quarterly 73 (winter 2009): 679–703. 23. Alan I. Abramowitz and Kyle L. Saunders, “Is Polarization a Myth?” Journal of Politics 70 (April 2008): 542–55; Marc J. Hetherington, “Resurgent Mass Partisanship: The Role of Elite Polarization,” American Political Science Review 95 (September 2001): 619–31; Gary C. Jacobson, “Partisan and Ideological Polarization in the California Electorate,” State Politics and Policy Quarterly 4 (June 2004): 113–39; Gary C. Jacobson, “Polarized Politics and the 2004 Congressional and Presidential Elections,” Political Science Quar- terly 120 (summer 2005): 199–218. 24. Geoffrey C. Layman, Thomas M. Carsey, John C. Green, Richard Herrera, and Rosalyn Cooperman, “Party Polarization, Party Commitment, and Conflict Extension among American Party Activists,” American Political Science Review 104 (2, 2010): 324–46. 25. Fiorina and Abrams, “Political Polarization in the American Public.” 26. Ross K. Baker, “Political Cover Rules the Day: Lawmakers Rarely Work Together. They Merely Find Convenient Allies—i.e., Lobbyists—in Order to Get Anything Done,” USA Today, April 14, 2010, 10A; John Harwood, “Partisanship’s Influence On Health Bill’s Future,” New York Times, March 1, 2010, 18. McCluskey and Kim 17

27. Bennett and Iyengar, “New Era of Minimal Effects?”; Young Mie Kim, “Issue Publics in the New Information Environment: Selectivity, Domain-Specificity, and Extremity,” Communication Research 36 (April 2009): 254–84. 28. Harwood, “Partisanship’s Influence.” 29. Baker, “Political Cover.” 30. Abramowitz and Saunders, “Is Polarization a Myth?”; Morris P. Fiorina, Samuel J. Abrams, and Jeremy C. Pope, “Polarization in the American Public: Misconceptions and Misreadings,” Journal of Politics 70 (April 2008): 556–60; Treier and Hillygus, “Nature of Political Ideology.” 31. Fiorina and Abrams, “Political Polarization in the American Public”; Fiorina, Abrams, and Pope, “Polarization in the American Public.” 32. Mark D. Brewer, “The Rise of Partisanship and the Expansion of Partisan Conflict within the American Electorate,” Political Research Quarterly 58 (June 2005): 219–29. 33. Fiorina and Abrams, “Political Polarization in the American Public.” 34. Brewer, “Rise of Partisanship”; Fiorina and Abrams, “Political Polarization in the American Public.” 35. Abramowitz and Saunders, “Is Polarization a Myth?” 36. Layman et al., “Party Polarization, Party Commitment.” 37. Abramowitz and Saunders, “Is Polarization a Myth?” 38. Alan I. Abramowitz, The Disappearing Center: Engaged Citizens, Polarization, and American Democracy (New Haven, CT: Yale University Press, 2010). 39. Fiorina and Abrams, “Political Polarization in the American Public.” 40. Abramowitz and Saunders, “Is Polarization a Myth?” 41. Morris Fiorina, Samuel Abrams, and Jeremy Pope, Culture War? (New York: Pearson and Longman, 2006). 42. Jonathan Rauch, Government’s End: Why Washington Stopped Working (New York: Pub- lic Affairs, 1999); Kay Lehman Schlozman and John T. Tierney, Organized Interest and American Democracy (New York: Harper & Row, 1986); Jack Walker, Mobilizing Interest Groups in America (Ann Arbor: University of Michigan Press, 1991). 43. Baker, “Political Cover.” 44. Fiorina, Abrams, and Pope, Culture War? 45. Fiorina, Abrams, and Pope, Culture War?, p. 8. 46. Bennett and Iyengar, “New Era of Minimal Effects?” 47. Bennett and Iyengar, “New Era of Minimal Effects?” 48. Shanto Iyengar and Kyu S. Hahn, “Red Media, Blue Media: Evidence of Ideological Selec- tivity in Media Use,” Journal of Communication 59 (1, 2009): 19–39; Kim, “Issue Pub- lics”; Silvia Knobloch-Westerwick and Jingbo Meng, “Looking the Other Way: Selective Exposure to Attitude-Consistent and Counterattitudinal Political Information,” Communi- cation Research 36 (June 2009): 426–48. 49. Albert C. Gunther, Cindy T. Christen, Janice L. Liebhart, and Stella Chih-Yun Chia, “Con- genial Public, Contrary Press, and Biased Estimates of the Climate of Opinion,” Public Opinion Quarterly 65 (3, 2001): 295–320. 50. Pew Center, accessed October 12, 2010, at http://people-press.org/report/652/. 18 Journalism & Mass Communication Quarterly XX(X)

51. Bennett and Iyengar, “New Era of Minimal Effects?” 52. Smerconish, “Media’s Black-and-White World,” A17. 53. For example, Lucig H. Danielian and Benjamin I. Page, “The Heavenly Chorus: Inter- est Group Voices on TV News,” American Journal of Political Science 38 (November 1994): 1056–78; Young Mie Kim and Michael McCluskey, “Revisiting the ‘Myth of Outside Strategies’: Information Technology, Group Strategies, and News Media Cov- erage” (paper, American Political Science Association conference, Toronto, September 3–6, 2009); Michael McCluskey, “Activist Group Attributes and Their Influences on News Portrayal,” Journalism & Mass Communication Quarterly 85 (winter 2008): 769–84; A. Trevor Thrall, “The Myth of the Outside Strategy: Mass Media News Coverage of Inter- est Groups,” Political Communication 23 (4, 2006): 407–20; Anke Tresch, “Politicians in the Media: Determinants of Legislators’ Presence and Prominence in Swiss Newspapers,” International Journal of Press/Politics 14 (January 2009): 67–90; Anke Tresch and Man- uel Fischer, “Political Actors in Search of Media Attention: An Analysis of Mobilization and Communication Strategies in Seven European Countries” (paper, American Political Science Association meeting, Boston, August 28–31, 2008). 54. William A. Gamson and Gadi Wolfsfeld, “Movements and Media as Interacting Systems,” Annals of the American Academy of Political and Social Science 528 (1993): 114–25. 55. Tresch, “Politicians in the Media.” 56. Tresch, “Politicians in the Media”; Gadi Wolfsfeld, Media and Political Conflict. News from the Middle East (Cambridge, UK: Cambridge University Press, 1997). 57. Paul Lazarsfeld and Robert Merton, “Mass Communication, Popular Taste and Organized Social Action,” in Mass Communications, ed. Wilbur Schramm (1948; repr., Urbana: Uni- versity of Illinois Press, 1960), 492–512. 58. Perhaps the first to use symbolic annihilation in referring to the absence of coverage was George Gerbner, “Violence in a Television Drama: Trends and Symbolic Functions,” in Television and Social Behavior, Reports and Papers, vol. 1: Media Content and Control. A Technical Report to the Surgeon General’s Scientific Advisory Committee on Televi- sion and Social Behavior, ed. George A. Comstock and Eli A. Rubinstein (Rockville, MD: National Institute of Mental Health, 1972), 28–187. Condemnation and trivializa- tion, along with absence, were included in the conceptualization by Gaye Tuchman, “The Symbolic Annihilation of Women by the Mass Media,” in Hearth and Home: Images of Women in the Mass Media, ed. Gaye Tuchman, Arlene Kaplan Daniels, and James Benet (New York: Oxford University Press, 1978), 3–38. 59. Gamson and Wolfsfeld, “Movements and Media.” 60. Bernadette Barker-Plummer, “Producing Public Voice: Resource Mobilization and Media Access in the National Organization for Women,” Journalism & Mass Communication Quarterly 79 (spring 2002): 180–205; Youngnin Yoon, “Examining Journalists’ Percep- tions and News Coverage of Stem Cell and Cloning,” Journalism & Mass Communication Quarterly 82 (summer 2005): 281–300. 61. Michael Gurevitch and Mark R. Levy, “Introduction,” in Mass Communication Review Yearbook, vol. 5, ed. Michael Gurevitch and Mark R. Levy (Beverly Hills, CA: Sage, 1985), 11–22; Charlotte Ryan, Kevin M. Carragee, and Cassie Schwerner, “Media, Movements, McCluskey and Kim 19

and the Quest for Social Justice,” Journal of Applied Communication Research 26 (May 1998): 165–81. 62. See Dave D’Alessio and Mike Allen, “Media Bias in Presidential Elections: A Meta-Analysis,” Journal of Communication 50 (December 2000): 133–56. 63. Based on the RR1 formula suggested by the American Association for Public Opinion Research. 64. Unlike individuals’ self-report of ideological leaning, groups’ ideological leaning is much more explicit and less biased by social desirability, since ideology is usually clearly stated publicly in mission statements. To validate the measure of ideology, we checked about 7% of the samples (randomly drawn) and matched their self-report with their mission statement and activities on their websites. No reverse-matching or mismatching cases were found. 65. BurrellesLuce, 2007, accessed at http://www.burrellesluce.com/top100/2007_Top _100List.pdf. 66. The very conservative and very liberal groups were compared on each presence variable analyzed and were significantly different on just one, as the very liberal groups were mentioned in newspapers with a significantly larger circulation than those that mentioned the very conservative groups. In addition, the polarized versus moderate groups were con- structed in various ways from the 7-point ideological scale (e.g., very conservative/very liberal/conservative/liberal as polarized and moderate/lean conservative/lean liberal as moderate) with substantially similar results to those presented here. Some additional differences in tone are evaluated in the results section. 67. Fiorina and Abrams, “Political Polarization in the American Public”; Fiorina, Abrams, and Pope, Culture War? 68. Our measure was conceptually similar to that of Schiffer, “Explaining Policy Coverage,” who created a moderatism/extremism scale from the Americans for Democratic Action Liberalism Quotient. 69. Editor & Publisher International Year Book (New York: Editor & Publisher Co., 2006). 70. Schiffer, “Explaining Policy Coverage.” 71. Croteau and Hoynes, By Invitation Only. 72. Kuklinski and Sigelman, “When Objectivity Is Not Objective.” 73. Gans, Deciding What’s News, 41. 74. Bennett and Iyengar, “New Era of Minimal Effects?” 75. Fiorina, Abrams, and Pope, Culture War?, 3. 76. Smerconish, “Media’s Black-and-White World,” A17. 77. Fiorina, Abrams, and Pope, Culture War? 78. Timothy Cook, Making Laws and Making News (Washington, DC: Brookings Institution, 1989), 49. 79. Michael Schudson, The Power of News (Cambridge, MA: Harvard University Press, 1995). 80. Lazarsfeld and Merton, “Mass Communication.” 81. Elmar Schattschneider, The Semisovereign People: A Realist’s View of Democracy in America (New York: Holt, Rinehart & Winston, 1960). 82. Gerbner, “Violence in a Television Drama”; Tuchman, “Symbolic Annihilation of Women.” 20 Journalism & Mass Communication Quarterly XX(X)

83. A focus on presence, however, carries with it some limitations. For instance, presence alone cannot capture the way that the groups and their goals are portrayed, which could influence public perceptions. Still, advocacy groups in the news have a better chance to influence public perception and policy concerns than those absent from the news. Even a less than positive portrayal of a group may be preferable to absence. Or, in the words of Todd Gitlin, The Whole World Is Watching: Mass Media and the Making and Unmaking of the New Left (Berkeley: University of California Press, 1980), 3, “Mass media define the public significance of movement events or, by blanking them out, actively deprive them of larger significance.” Public perception, in turn, could fuel broad public perceptions of polarization and among some individuals in which the polarizing effect of media may be particularly strong; see Cathy F. Bullock, Michael McCluskey, Keith Stamm, Keiko Tanaka, Marcos Torres, and Cathie Scott, “Group Affiliations, Opinion Polarization, and Global Organizations: Views of the World Trade Organization before and after Seattle,” Mass Communication & Society 5 (4, 2002): 433–50.