Disinterested white knights: On the sale of U.S. newspapers to private individuals
and how a change in ownership affects political coverage and interpretive journalism
Edmund Austin Ruff IV Student ID: 12367648
Master’s Thesis
Graduate School of Communication University of Amsterdam
Master’s program Communication Science Supervisor: Joost van Spanje
Abstract
This paper looks at recent cases of major U.S. newspapers being purchased by wealthy, private individuals who come from outside the media industry. Theory suggests that journalism output is subject to influences and exchanges between journalists and the owners/investors who employ them, and that individual owners, compared to corporate owners, will the most control over their media. Previous research has stated that different types of ownership effect different types of news coverage, suggesting private owners are more ideologically driven but politically neutral. But little work has taken a before-and-after approach to ownership changes, to explore what happens over time to a newspaper’s content, in terms of coverage and quality.
This paper, then, looks at two newspapers – The Washington Post and the Star Tribune – that were purchased by private individuals. Using a before-and-after design, front-page articles from pre-sale and post-sale periods at each newspaper were analyzed to measure shifts in their respective amounts of political coverage and ideological news quality, conceptualized as interpretive journalism. Those changes were compared to competing newspapers with static ownership to show relative change.
In line with prior literature, it was expected to find an increase in both the amount of political coverage and interpretive journalism at the newspapers that changed ownership. However, the Post and the Tribune showed little change in relation to their market competitors, suggesting the measured changes cannot be attributed solely to their change in ownership. Surprisingly, the one newspaper that showed the most change had static corporate ownership.
A counterpoint to previous research on ownership influences, this study provides evidence that journalism output is not so easily manipulated under private ownership.
Introduction
When the sale of The Washington Post to Amazon.com co-founder and CEO Jeff Bezos was announced in 2013, cautious optimism followed. An news industry outsider and one of the world’s wealthiest people, Bezos brought a “willingness to finance experimentation” (“Exploring the Amazon”, 2015) at the newspaper, which had been hobbled by years of cuts under its
previous owners. The Post’s staff had been reduced, and its circulation had been sliced in half from its peak in 1993 of 832,000, to 432,000 in 2015 (Kennedy, 2018). According to the Alliance for Audited Media, that decline in circulation was larger than the national average decline over the same period (“Paid Circulation of Daily Newspapers in the U.S. 2016”, 2018). Bezos’ purchase of the paper was seen as a lifeline to the newspaper.
Bezos’ purchase of the Post is just one example of a recent trend of newspapers sold from corporate media companies to private, wealthy individuals. An early (if inexact) example of this was the sale of The Wall Street Journal (WSJ) in 2007, to Rupert Murdoch’s News Corp. In the 2010s, the sale of newspapers to other wealthy individuals became a trend. Bezos’ acquisition of the Post was soon followed by the sale of The Boston Globe in 2013 to John Henry, a
businessman and investor; the Minneapolis Star Tribune in 2014 to Glen Taylor, CEO of a multinational electronics company; the Las Vegas Review-Journal in 2015 to Sheldon Adelson, owner of Las Vegas Sands casino company; and the Los Angeles Times in 2018 to Patrick Soon-Shiong, a biotech investor. Notably, all the buyers of these newspapers are persons who made their fortunes outside the media and journalism industry. What’s more, all of these newspapers are among the 25-largest in the U.S. by circulation1.
This trend of billionaire industry outsiders stepping up as “white knights” (Herman & Chomsky, 2002) to provide capital to foundering newspapers raises questions about ownership and media autonomy. Of particular interest is what these industry outsiders – with virtually no
experience in news media production – want to do with the newspapers they have purchased; do they intend to restrict or enhance the news and content produced in order to turn
once-floundering print media into profitable outlets? Or will they do nothing but provide funding to keep the business solvent? Previous authors call the latter outcome unlikely. “All capital comes with conditions. … In private ownership those who provide the capital have a high degree of control” (Picard & van Weezel, 2008, p. 23). This is not to say owners exhibit direct control over newspaper production. But the hierarchy of influences model (Reese & Shoemaker, 2016) suggests there are layers of stimuli, both explicit and implicit, on journalists and their output. “The incentives to create coverage that is favorable to the interests of the media owners is likely a result of journalists’ incentives to weigh the preferences of their profit-maximizing owners” (Bailard, 2016, p. 589). Theory suggests, then, that a change in ownership will lead to a change in influences on those who produce news and, subsequently, a change in the news produced. How that change manifests in news content remains largely unexplored. This leads to the main focus on this study:
RQ: When a newspaper changes to private ownership, how does that change the news content produced?
This research is important for answering ongoing questions about the future of news production. As the digital age progresses, traditional media have sought new funding models to remain viable. Print media in particular have faced an existential crisis after years of declining circulation numbers, consolidation and corporatization. Amid a landscape where a “handful of dominant corporations … pursued quick, ever higher profits” while pursuing journalism that would “promote the needs of the owning corporation” (Bagdikian, 2000, p. ix-xii), the new cases of private ownership by industry outsiders – away from conglomerate and publicly-traded corporate ownership – represent a shift in possibilities for the future of journalism.
Beyond answering industry-specific questions of financial sustainability, the question of who owns the news and how their ownership affects news content matters for democratic societies as a whole. Quality journalism is vital to quality democracy. “Media have become the most important source of information for most people in advanced democracies around the world” (Strömbäck, 2008, p. 229). Further, a necessary component of a successful democracy is an informed and active citizenry; being informed means knowing where information originates, through which filters it passes and under what influences it was created.
On newspapers sold to wealthy industry outsiders, little research has examined how this “white knight” form of funding journalism affects the content produced. What research there is takes a snapshot of a singular points in time, with results that suggest corporate ownership produces less substantive and less neutral news in relation to newspapers under private ownership (Dunaway, 2008, 2013), while other research (Hanretty, 2014) suggests private owners are perceived as having a higher degree of influence on content. These studies fail to show change over time and how a change in newspaper ownership might affect news content. Two studies that have taken change over time into consideration suggest a distinct increase in political content (Archer and Clinton, 2017; Wagner and Collins 2014), but their focus is mainly on opinion content rather than news content. Further, their focus is on one newspaper in
particular, the WSJ, an early case that – because of Murdoch’s established place within the media industry – is not fully comparable to the more recent examples2. However, these prior
studies’ focus on political news content lays foundation for this paper.
Studying the effects of a change in ownership on news content is vital to understanding the what impact private owners might have on news consumers. Whoever owns the news outlet has direct influence on the news content, and thus over the information and opinions given to consumers. However, current theory focuses explicitly on corporate and conglomerate news
media; therefore, current scientific knowledge lacks precision in predicting the effects of private individuals buying up major newspapers. This paper aims to add to the literature on media ownership and bring attention to changes in the industry.
Theory
This thesis is informed by three models of the influences on news production. First, the market-based model of news production proposed by McManus (1995) illustrates that, within the production cycle, a series of interconnected markets determines the news output, as the various players engage in an exchange of resources, such as information, attention, influence and capital. At the top of the model, investors and owners engage in a top-down exchange with media
corporations, exchanging capital for profits and influence. McManus explicitly sets his model in “Western-style societies where private businesses are the primary news providers” (p. 305). The model, however, does not account for singular ownership of media, instead defining owners as the broad range of stockholders of large conglomerate media. Despite dissonance in application to this thesis, the market-based model does anticipate broad “information power” given to those with most financial stake in the news company; “the greater the proportion of a company owned by a stock holder, the more influence that person has” (p. 307) on the news output.
A second model this thesis draws on is the hierarchy of influences outlined by Reese and Shoemaker (2016). They identify five levels of analysis that shape news content, from the
individual (micro) level to the societal (macro) level, further describing how news is a product of individual agency within set media structures. Individuals who produce news have their own agendas that are “both constrained and enabled by the structures surrounding them” (p. 397). The implication is that changing the organizational structure in which an individual operates could effect change on the news that that individual produces; more simply, new owners means new
influences. This represents a top-down approach, with the individual at the bottom is influenced by the above organizational structures, including the owner.
A final theory that, in part, informs this thesis is the propaganda model by Herman and Chomsky, which describes five filters that affect news output. The first of these filters,
ownership and profit orientation, states that news that is counter to owners’ interests will not make it past the first filter without censorship or imparted bias. Individual journalists lack full autonomy as they work to uphold the interests of their employers. The propaganda model, however, emphasizes this filter process as adhering to the interests of corporate or conglomerate media owners, making it imprecise (and outdated) in regard to the current phenomenon of interest, namely that of individual and private ownership.
These theories outline the mechanism of interest for this study, that news output is a direct result of ownership type; different types of ownership, therefore, are expected to bring different influences. As the providers of capital, media owners have broad control over their media organizations, manipulating the output both explicitly and implicitly. Hanretty
summarized these influences as those “exercised overtly, through company memos, irate phone calls or interference in hiring and dismissal,” while also noting that “influential owners rarely have to make their wishes known in order to have their will done” (Hanretty, 2014: p. 338).
For this paper, the focus of ownership is on private owners from outside the media industry. Because they have made their capital outside the journalist industry, they have less interest in getting immediate profit from the newspaper, and therefore might take a “hands-off” approach to media production, meaning less influence. Compare that to industry insider
ownership, like at the WSJ, where Murdoch has been observed to take a “hands-on” approach to influencing media content to ensure profit. As another comparison, corporate ownership relies on profit-seeking shareholders who have minimal direct influence – or, rather, highly diffuse
influence – on the news product.
Regarding changes to news content, private industry outsider owners could go down one of three paths when taking over a newspaper, as suggested by the market-based model of news production. First, because industry-outsider owners are less interested in profits from the news business, they could encourage more ideological journalism, at the risk of alienating some readers who disagree with the ideologies and interpretations the newspaper produces. Second, the owners could go the other way and discourage interpretive journalism because of its potential to damage the "objectivity" standard. A third possibility is the fully disinterested hands-off owner, who lets the journalist and editors produce news with little to no top-down interference.
Before proceeding, it is important to note that all for-profit media seek profit and growth, and individual owners especially should want a return on investment. The distinction among different types of private and corporate ownership is that “private owners have other competing goals, such as journalistic goals, which may compete with the profit making objective” in the short-term; this is far and away different from the “constant effort at profit maximization” sought by shareholders in the quarterly or yearly returns at corporate media (Dunaway, 2013: p. 27).
Political journalism
Thus far, news content has been used as a catch-all concept for all non-advertising material that is written for and published in a newspaper. In practice, though, news content can manifest in any number of forms and with any kind of subject matter. To assess news quality for this paper, there is a need to narrow the focus on content to avoid getting lost in the woods. So far, previous studies that have explored the effects of ownership on news content have all focused specifically on political coverage. Why? A change in sports coverage could lead to a change in readership and circulation revenue. Likewise, promotional business coverage could inspire more advertising revenue. Owners exchange capital for a return on investment, according
to McManus’ market-based model; one would expect, therefore, that studies of ownership influences would focus on areas that the owners might want to expand to maximize their profits.
Dunaway provided justification for this focus on political news, stating that “the primary purpose and most important function of the media … is to provide the public with sufficient information for evaluating leaders and governance” (2008: p. 1193). Put another way by the Pew Research Center, news media’s essential mission is to provide “the news and information [that people] need to be effective citizens” (Rohal, 2013). To that end, the clearest form of news content by which to analyze citizen- and governance-oriented news is that of political coverage. To be considered quality, a newspaper must contain political coverage. The amount of political coverage, then, is tied to indicators of quality.
Politics at all levels is inherently interlinked with subjective ideologies – both implicit and explicit – that focus on personalities and principles, as opposed to objective facts that define daily events coverage. Further, perceived news quality is more fluid in political stories. This fluidity allows for a more targeted approach in the search for actual changes in how news content manifests. It is assumed that different news outlets will cover objective news in much the same way, while subjective content in political news should elicit more clear differences from one outlet to the next and, indeed, within the same outlet over time and alongside a change in ownership.
Interpretive journalism
The establishment of a comparative line of thought on changes in news media ownership has been described as “one of the most fundamental needs” in the field of political
communication research (Hallin & Mancini, 2012, p. 218). Taken a step further, some have sought to delineate a standardized meaning of “good news” and why some media systems produce more of it than others (de Vreese, Esser & Hopmann, 2017). One of the factors they
identify for comparing political news is the appearance of interpretive journalism. Patterson characterized this type of news content as “facts and interpretations [that] are freely intermixed” (1993, p. 67). Using the term “long journalism”, Barnhust and Mutz described it as a shift “from episodic to thematic coverage”, where reporters “deemphasize events in favor of news analysis” (1997, pp. 27-28). Salgado, et al., called interpretive journalism “one of the key trends of
journalism during the last decades” (2017, p. 50).
Interpretive journalism was conceptualized early on as agenda-setting journalism (Blumler & Gurevitch, 1975), where decisions about news coverage were driven by journalists more than by the news events and sources. The study of interpretive journalism did not gather steam until the mid-1990s and 2000s, where a distinction was made “between a mainly
descriptive and a mainly interpretive journalistic style” (Salgado & Strömbäck, 2011, p. 148). In short, interpretive journalism is a shift from a multifaceted balance of “who, what, when, where and why” to a singular focus on an overweight “why”. The study of interpretive journalism is important for political news, as it marks a reliance on the journalist or author to determine the meaning of news, delegating the reader a more passive role in news consumption – a passivity in direct contrast to democracy’s need for an informed and active citizenry.
The search for quality news standards is not a necessarily normative judgment; “good” news is relative to an outlet’s market and audience and to the topic of the news itself. Because newspapers in democratic societies are charged with keeping the citizenry informed, high-quality news must allow readers to craft their own informed opinions about even the most complex events. Particularly for political news, the use of interpretive journalism is one tool to aid readers in the opinion-making task, and high-quality news is expected to include explanations to expand the scope of an article’s relevance and allow the reader to synthesize the meaning of events with their own ideas. Too much interpretation, however, can impair the reader’s ability to craft
personal opinions, creating a passive citizenry that absorbs pre-fabricated views direct from the news outlet. It is crucial, then, to examine differences in the appearance of interpretive
journalism at different news outlets, to illuminate differences in political news quality.
Literature review
Previous studies have established that different types of ownership have different
influences. Until recently, most studies looked at concentration of ownership at the market level – how many news outlets one company controls. Rejecting categorical strokes and taking a quantitative approach, Hanretty (2014) looked at the concentration of ownership at the outlet level, that is, how much of an outlet is owned by how many people. Looking at 211 print and broadcast outlets throughout the European market media, Hanretty concluded that closely held media outlets – those with higher proportions owned by fewer individuals – were perceived as functioning under more influence from those owners than outlets with diffuse or widely held ownership. This study, however, relied on a survey of experts and their perceptions and did not look at content quality at those media companies. The author went so far as to criticize content analysis as “conceptually flawed” (p. 338) in identifying differences in news output across ownership levels. Nevertheless, Hanretty’s work is important in establishing that, while news outlets with different ownership types do not necessarily produce news differently, they can be
perceived by audiences though they do.
In a pair of studies, Dunaway (2008, 2013) brought the issue of news content differences between private vs. corporate newspaper ownership into focus. The first study examined print and television coverage of two 2004 elections in both Colorado and Washington state; although the selection of narrow media markets risked generalizability of the findings, the study made note of exogenous forces, such as market competition, in effort to stabilize any potential
differences between the sample and the broader industry. The coverage was rated in terms of events-based news and issues- or substance-based news, which Dunaway deemed as news of lower and higher quality, respectively. The results determined that private-owned media produced more issues-based news on candidates’ platforms, compared to corporate-owned media, which reported more campaign events and horse-race-style election coverage. The author categorized private owners as ideological, meaning they were “more likely to prefer substantive issue coverage” on specific issues the owners think important; this was compared to corporate media, which aimed for “broad audience appeal” with its profit-motivated production of less substantive news (2008, p. 1195). Dunaway was careful not to position owners’ ideologies as politically anchored or corporate mass appeal as inherently objective; verily, ideology can manifest as politically neutral, and mass appeal can lead to distinct bias.
Dunaway’s 2013 study took aim at clarifying ownership type and news bias in terms of story tone at newspapers with different types of owners during competitive state-level elections. Although it focused on local and smaller newspapers with static ownership, it drew its sample from nearly two dozen states, enhancing its generalizability across the U.S. Its findings
determined corporate-owned newspapers produced more skewed news coverage, while private-owned newspapers produced more balanced or neutral coverage. The study’s reasoning was that non-neutral news fit corporate media’s profit motives by “attracting and retaining audiences” by use of alternate voices (p. 40). Taken together, Dunaway’s two studies suggest private ownership results in balanced news coverage with more substance, while corporate ownership tends to eschew policy issues but favor slanted or skewed news content.
These studies establish the perceived and demonstrated differences in content at media outlets with different ownership types, but they do little in showing change over time, relying instead on static snapshots to show differences in content at singular times.
Changes in ownership and political content
In one of the first studies to look at before-and-after changes in news content with an eye on ideological owners, Wagner and Collins (2014) dove into Murdoch’s purchase of the WSJ. Rather than news articles, the authors specifically examined news editorials – unsigned opinion pieces typically produced by a newsroom’s lead editors, a sampling decision the authors justified because editorials typically have “a major influence on content of reporting … and public
evaluations of political issues” (p. 759). In comparing the WSJ to two competitors, The New York Times and The Washington Times, the study found that, while each newspaper saw a shift in ideological slant – suggesting that ownership is not the sole influence on editorial content – the WSJ exhibited the largest shift to the political right after Murdoch’s takeover. The authors admit this result was not altogether unexpected, as prior research had explored how Murdoch’s News Corp. “carefully directed the policy orientation of [its] other media outlets” (p. 768); moreover, owners have broad control on opinion-based editorials and a change in their ideological focus after a change in ownership should not be surprising.
The WSJ’s ownership change was the focus of a separate study, by Archer and Clinton (2017), who examined editorials and front-page content at the WSJ compared to The New York Times, USA Today, and The Washington Post. While not contradicting Wagner and Collins’ focus on ideological slant in the editorials, Archer and Clinton did not find a stark change in the number of politics-focused editorials over the pre- and post-Murdoch years nor compared to the other papers. Instead, they attributed the changes to the sampling periods, namely presidential election years. However, their data on front-page content found that, while each of the
newspapers they studied did increase political news coverage, the WSJ increased its political coverage at a rate much higher than the others, a 17.24 percentage-point increase, compared to the New York Times (6.76), USA Today (3.45) and the Post (1.93). Critically, Archer and
Clinton do not address that the WSJ is a business and economics newspaper, which covers topics that, while not divorced from politics, tend to operate separately from government. In essence, the WSJ had room to grow when counting political articles, where the other newspapers had far less room to grow. Further, this study’s sample period ended with 2008, the height of the financial crisis, a global event that was both economically and politically rooted. The study did not include any data from after the crisis began, which would have shown whether the WSJ returned to its lower levels of political coverage. Though not without flaws, these results show the effect that ownership can have on a media content, crucially laying the foundation for explaining how a change in ownership might affect news content.
Because news is subject to a variety of influences, it is anticipated that a change in political content alongside a change in ownership might be more influenced by the actual news of the day. There is no standard for how much political news a newspaper should carry, but that amount is determined by innumerable temporal factors. As Bailard asserts, the absence of an “objective and absolute standard” (2016, p. 585) can be rectified, by standardizing news factors as relative to themselves – that is, the appearance of coverage at one news outlet in relation to the appearance of coverage at other, similar outlets.
From this perspective, an initial hypothesis is drawn to illuminate whether any changes in the amount of political news at a given newspaper are a result of its change (or lack thereof) in ownership rather than a simple change in the “news of the day.”
H1: A newspaper will increase its amount of political coverage after being sold to a
private owner, compared to newspapers with stable ownership in its same market. Changes in interpretive journalism
The increase of interpretive journalism in recent decades has been termed “the most significant change in political journalism” (Djerf-Pierre & Weibull, 2008, p. 209). Salgado &
Strömbäck (2011) characterize interpretive journalism as the intermingling of facts and themes, rearranging the five W’s of reporting – who, what, when, where and why – to place more emphasis on why.
While journalists frequently cannot overtly question what political news sources say, without being perceived of being biased, they can decide on the thematic framework within which the actions and words of political news sources are placed, and analyze the meaning of the What, Where, When and Who. (p. 154)
Interpretive journalism, therefore, increases the role of the journalist in adapting information from sources to craft a larger narrative. This is opposed to descriptive journalism, where the journalist merely transfers information from sources largely intact.
In their definition, Salgado and Strömbäck highlight five indicators of interpretive journalism: prominent journalistic voice, explicit labeling of analysis, evaluative commentary, unsourced explanations, and speculation. These are qualified as “going beyond verifiable facts or statements by sources” (p. 154). Further, they highlight that, through different combinations of the indicators, it would be possible to identify different types of interpretive journalism, but they do not go so far as to delineate a typology. This intentional omission, they stress, is to avoid a normative assessment of interpretive journalism, which would suggest one type of journalism being “better” than another. This, too, is a recognition of the absence of an objective standard in this area, “as the assessment of different kinds of interpretive journalism may vary” (p. 155), especially depending on the media type and the exact news outlet. Market factors are at play, as different audiences in different locations would expect different amounts of interpretive
journalism from their chosen news outlets. However, by controlling for similarities across markets, it might be possible to remove variance in the assessment of interpretive journalism from one outlet to the next.
ownership poses stark questions for how meaning is determined, whether from a singular
influence presenting a singular meaning or a diffuse influence allowing for more varied meaning. Prior research has explored how different types of owners can be placed on an ideological
spectrum, suggesting private, individual owners are more focused on a specific set of ideologies and meanings compared to corporate owners. As such, it is predicted that the amount of
interpretive journalism will follow the type of media owner, and that a change to private
ownership will lead to an increase in interpretive journalism. Again, any changes at a newspaper under new ownership must be standardized in relation to other newspapers to control for “news of the day” influences.
H2: A newspaper will have more interpretive journalism after being sold to a private
owner, compared to newspapers with stable ownership in its same market
Methodology
The focus of this paper is on newspapers that changed to private ownership and what changes appear in content over time. For this study, two newspapers were targeted for analysis: The Washington Post and the Minneapolis Star Tribune.
Rather than editorial-page content, like previous studies have done, it was determined to use only front-page news articles. This follows Archer and Clinton, who assert that front-page news is most likely to be affected by real-world events, which should “illuminate the nature of the change in focus that may result from a change in ownership and emphasis” (2017, p. 5). This assumption indicates that the articles on the front page would be the most independent of other influences – be they managerial or philosophical – that would affect a newspaper’s interior, such as number of pages or number of advertisements. In short, all printed newspapers regardless of ownership type have a front page. Keying in on the single page most free from other pressures should signal whether a change in news content is a result of a change in ownership, as opposed
to some other influence. Because a newspaper can carry a different number of articles on each day’s front page, the samples are expected to be non-proportionate.
A stratified sample was selected to fit with the before-and-after theme on ownership change, creating a comparison of one newspaper with itself. First, it was necessary to establish a post-sale sample for both newspapers. Because the Tribune was sold last, the post-sale sample was established on that sale date. A delay of one year from the Tribune’s sale date was chosen to allow any ownership influences to take effect, under the assumption that any change would not be immediately apparent. Based on that timeline, the post-sale sample was a 28-day period from 30 June to 27 July 2015. For the pre-sale sample, any 28-day period that occurred before the sale of the Post would work. Because of H1’s focus on political coverage, the sampling strategy was to follow the U.S. four-year presidential election cycle, to make the seasonal political themes as similar as possible. Therefore, the pre-sale sample dates were 30 June to 27 July 2011. These “off years” in the election cycle are ideal to study general political coverage to reduce the effects of the juggernaut that presidential-election coverage typically is.
The search for a relative standard of news quality demanded a wider selection of newspapers than just the Post and the Tribune. If the study were limited to examining two newspapers that changed ownership, there would be no baseline by which to compare any calculated changes against newspapers that did not change. Therefore, each newspaper was paired with a newspaper from its same geographical market that had static ownership over the chosen sampling periods. The Tribune, published in Minnesota’s Twin Cities region, was paired with the St. Paul Pioneer Press. While the comparison was not perfect, because the Tribune focuses more on state/regional news and the Pioneer-Press delivers a wider variety of local, national and international news, the two newspapers offered a close comparison on their reporting of state government news.
For the Post, published in the Washington D.C. region, the initial partner was USA Today. After pulling the initial sample of USA Today, it was decided to add The New York Times as well. USA Today targets a broader readership than that of the Post and the Times, which, although published in different cities, are considered more competitive on the national market. The difference is evidenced by the sample numbers, which revealed the Post carried an average of 5.1 articles on the front page per day, compared to the Times’ 5.9 and USA Today’s 1.9. (The Tribune carried 3.3 front-page articles per day, compared to the Pioneer Press’ 4.2.)
Articles for the Post, the Tribune, USA Today, and the Times were collected using the Nexis Uni archive. The Pioneer Press articles were collected using the newspaper’s own online archive3. Keywords used to tag the article as front page differed for each newspaper; for example
the Post articles were tagged “A01”, while the Tribune articles were tagged “1A”. Only full articles were collected; teasers – articles of two or three sentences that previewed full-length articles on the inside pages – were excluded from the sample.
The final corpus was 285 articles for the Post, 106 for USA Today, 328 for the Times, 186 for the Tribune, and 237 for the Pioneer Press – a total of 1,142 articles.
Because this paper hinged on comparison within a newspaper to find change over time that occurs alongside change in ownership, each article downloaded from its respective archive was cleaned of extraneous detail, particularly its date of publication, to avoid imparting any influence on the coder. Each unit article contained only the given headline, the byline or author’s name, and the body of the article itself. The articles were then assigned an ID number using the random number function in Microsoft Excel. A master list of all the articles’ headlines,
publication dates and ID numbers were kept, to allow for a de-randomization after the coding was finished.
Despite a general idea of what constitutes politics, the literature disagreed on an operational definition. Other literature focused solely on elections coverage (Dunaway, 2008; 2013) while others defined politics as matters of “public policy” but concentrated on domestic or national-level policy (Archer & Clinton, 2017: p. 6). Given the variety of national-levels of politics, from
international agreements down to individual voters’ opinions, it was determined this paper would encompass all political issues. The definition used in the codebook was “anything related to the functions of government and the discussion of policy at the local, state, national and international levels (and every step in between), including but not limited to elections and referenda, citizens' initiatives, the actions of political parties and individual politicians, judicial interpretations of statutes at the U.S. or state Supreme Court levels (not jury trials or criminal court rulings), and international diplomacy.” (For the full codebook, see Appendix.)
The first question in the codebook asked whether the article was about politics. This functioned as a sort of filter for the corpus. If the article was coded “yes”, then coding continued. If “no”, the coding ended.4 After the “politics” filter was applied, 568 articles were subject to the entire codebook.
In developing a method to study interpretive journalism, some scholars have noted the mutability of its appearance in news publications. Indeed, no single news article can be called solely descriptive or solely interpretive. Such a dichotomous approach fails to recognize that even the hardest of facts “can also be the subject of an interpretive angle” (Salgado & Strömbäck, 2011: p. 154). To capture the nuance of interpretation, Salgado and Strömbäck proposed a standardized codebook that quantifies interpretive journalism to allow for the comparability of the different studies’ results. As they note, previous studies of interpretive journalism have been full of ambiguities that limit cross-comparison of their different methods. However, there has been limited adoption of their codebook. This paper was envisioned, then, as
an attempt to test and refine their codebook to fit their vision of conceptually clear comparative studies of interpretive journalism. Their codebook consisted of seven indicators:
- presence of explicit labeling of interpretive journalism (categorical/binary) - presence of journalistic explanations (categorical/binary)
- presence of contextualization of events (categorical/binary) - presence of speculation about the future (categorical/binary) - presence of overt commentary (categorical/binary)
- overall amount of interpretations and explanations (ordinal) - amount of journalist’s voice (ratio)
After codebook testing revealed no article with any of the anticipated labels, the “explicit labeling” indicator was amended to whether the lead paragraph was considered analytical. The contextualization indicator was excluded from the analysis because it lacked the complete operational clarity found in the remaining categories. The ordinal measurement – which asked the “overall amount” of interpretive journalism to be labeled as high, medium or low – was found to be redundant next to the ratio measurement, which asked for the exact percentage of the journalist’s voice. The ratio measurement, however, was difficult to apply consistently, as the coding process revealed most articles were written in a manner that obscured the journalist’s voice among factually verifiable or quoted material. With the above considerations, the four remaining categorical indicators – analytical leads, unsourced explanations, overt commentary, and future speculation – were used in the final codebook and analysis5.
Inter-coder reliability tests using Krippendorf’s alpha (Hayes & Krippendorf, 2007) for each codebook question were conducted on a 10-percent sample for each newspaper. After numerous attempts, reliability for each question ranged from 0.0163 to 0.1920, indicating results only slightly better than random. The reliability test suggests the results have limited fidelity.
Results
In examining the data for H1, the two newspapers that changed ownership performed differently. The Post increased its proportion of political articles while the Tribune’s proportion decreased. Both newspapers’ changes were, however, in line with their market competitors (Graph 1). The Post and the Times were nearly identical in both years, with USA Today showing a small increase as well. The Tribune and the Pioneer Press both decreased. It was expected that newspapers changing to private ownership would increase their political coverage to exceed that of their competitors, but H1 is not supported by the data.
Table 1: Proportion of front-page political articles (%), compared within newspapers.
Graph 1: Proportion of front-page political articles (%) by year and newspaper.
Data set Washington Post
New York Times
USA Today Star Tribune Pioneer Press
N=281 N=327 N=106 N=185 N=237 Sample year 2011 2015 2011 2015 2011 2015 2011 2015 2011 2015 n 133 148 163 164 52 54 90 95 117 120 % of articles 49.62 58.78 48.47 58.54 44.23 48.15 50.00 32.63 52.99 44.17 (% change) (+18.46) (+20.78) (+8.86) (-34.74) (-16.64) n 66 87 79 96 23 26 45 31 62 53
For H2, the indicators of interpretive journalism were difficult to analyze. Salgado and Strömbäck, in their proposed codebook, did not detail a typology of what would indicate high or low interpretive journalism or explain how to combine their categorical and interval/ratio
indicators into a complete picture. For this paper, then, a simple categorizing of the data for H2 was conducted. All political articles that had at least one categorical indicator of interpretive journalism present were counted. Those numbers are summarized in Table 2 and Graph 2.
For H2, the data showed the same diverging performance between the Post and the
Table 2: Proportion of front-page political articles (%) with at least one interpretive journalism indicator present.
Graph 2: Proportion of front-page political articles (%) with at least one interpretive journalism indicator present, by year and newspaper.
Data set Washington Post
New York Times
USA Today Star Tribune Pioneer Press
N=153 N=175 N=49 N=76 N=115 Sample year 2011 2015 2011 2015 2011 2015 2011 2015 2011 2015 n 66 87 79 96 23 26 45 31 62 53 % of articles 87.88 98.85 94.94 94.79 56.52 73.08 86.67 70.97 82.26 83.02 (% change) (12.48) (-0.16) (29.3) (-36.48) (0.92) n 58 86 75 91 13 19 39 22 51 44 0 20 40 60 80 100
Post Times USA Tribune Press
%
Political articles with interpretive indicators
Tribune as for H1: The Post marked a significant increase in the proportion of articles with “at least one” indicator of interpretive journalism. This suggests journalists at the Post were more willing to use interpretive journalism after the newspaper’s sale. The Tribune, meanwhile, showed a steep decrease in the proportion of articles with “at least one” indicator.
The Post’s measurements, though increased, remained clustered around the Times. The Tribune’s decrease occurred while the Pioneer Press remained stable. Curiously, USA Today marked the sharpest increase, the sole newspaper with stable ownership to show any change. Although it was expected a newspaper changing to private ownership would measure more interpretive journalism compared to its competitors, the data does not support H2.
Discussion / Conclusions
This paper sought to explore how a change in newspaper ownership would affect content and quality. It was hypothesized that a change in ownership would create a newspaper with more political coverage, as previous research has shown happened at the WSJ, and that private
ownership would make a more ideological – or interpretive – newspaper, as previous research has shown occurs in small newspapers during competitive election campaigns.
As a whole, the data collected for this paper revealed little correlation to a change in ownership and a change in political news coverage. This refutes prior research on the increase of political coverage and the strengthening of ideology under private ownership, showing minimal change in either newspaper relative to their market competitors and suggesting the new owners took a “hands-off” approach. As a whole, the lack of support for either the hypothesis suggests a newspaper’s daily diet of front-page news is influenced more by a newspaper’s market and target audience than by its owner’s preferences.
political and less interpretive, especially compared to that of the Pioneer Press. This suggested its new owner wanted a less ideological newspaper. During coding, however, it became apparent the 2011 sampling period marked a unique political time for Minnesota; the four-week period saw the start and end of a state-wide government shutdown, as well as several profile articles on two Minnesota politicians who had announced high-profile campaigns for U.S. president. These types of articles were absent from the 2015 sample. A future study should look at a less politically vibrant era in Minnesota politics, to produce a cleaner year-to-year comparison.
It must be noted that, due to poor levels of inter-coder reliability, the results are tenuous at best. This study could have benefited from more time for coders to coordinate and harmonize the differing results. Further, the codebook was filled with difficult operationalizations of subjective concepts, proving Salgado and Strömbäck’s interpretive journalism codebook to be flawed and in need of expansion to capture unexpected nuances of language and meaning.
Given the significance of observed changes at USA Today, however, it is this author’s opinion that there is merit in the results. Because USA Today was an example of a newspaper under stable ownership, the changes were initially surprising. As the one newspaper that became markedly different between the two sample periods in relation to its competitors, the changes at USA Today hint at some influence beyond a change in ownership or the “news of the day”.
A post hoc review of USA Today’s recent history revealed it had undergone organizational changes between 2011 and 2015. Notably, the newspaper welcomed a new president and publisher, Larry Kramer, in May 2012 (Beaujon, 2012). According to an article published after Kramer stepped down in June 2015, his tenure saw the newspaper “transformed … from primarily a print publication to a digital news platform” (Snider, 2016). In the same time period, its print product underwent a major re-design as well as an overhaul of its digital
Co. Inc., called the new design “a signal of all the changes that are happening” and a
“re-imagination” of the newspaper’s offerings (Gosling, 2012). Taken together, the shift in political and interpretive focus at USA Today shows the changes at the newspaper were not merely cosmetic. Though USA Today’s shift in interpretive journalism reflects its new publisher, the Post also changed publishers in 2014. This suggests that internal leadership changes, like external changes, are one of many influences on news content.
This study was limited to just two newspapers that have changed owners in the past decade, providing a narrow look at what private industry-outsiders bring to the newspapers they buy. It is possible the chosen samples were narrow to allow time for changes to manifest at either newspaper. This could be addressed with a larger, longer sample from different years or different times of year, as well as looking at other areas of newspaper content that just front-page political content. Future research also should examine other newspapers – particularly those identified in the introduction – to explore how different individual owners might flex their influences in different markets.
Further, this study was limited to U.S. newspapers and made no attempt to study any outlet’s digital platforms. Particularly with the Post being added to Bezos’ Internet empire, future research could explore what might have manifested online instead of in print. Finally, because research has shown that private owners in non-U.S. settings are perceived as having a higher degree of control over their media outlets, work from a non-U.S. perspective on news content might clarify why this paper contradicts those findings.
At no point was a line drawn to declare any one newspaper in this study as high quality or low quality. Although tied to the amount of political coverage and interpretive journalism for this paper, quality is a broad concept that varies from one news consumer to the next. More research on different and more operationally clear measures of quality would identify how private and
corporate owners influence their media outlets.
Though this study was sparked by an interest in changing ownership of newspapers, the results suggest change comes not necessarily from an ownership change but mostly from the news of the day within a given market. In summation, if any changes are announced in the ownership of a newspaper, readers should watch for a change in news content, but need not necessarily expect it without considering the entirety of possible influences.
Endnotes
1 Rankings were based on circulations numbers reported by the Alliance for Audited Media’s March 2013 “Snapshot” report, which included a combined daily average of both print and digital subscriptions. Newspaper companies self-reported these numbers to the Alliance.
2 On the whole, the WSJ is not compatible with the other newspapers of identified in this paper. The WSJ was sold from one corporation to another, and though the new corporation is largely owned by Murdoch’s family, the family does not have majority ownership. Further, as some have noted, “News Corporation is an atypical news media corporation; it is sui generis [of its own kind]” (McKnight, 2010, p. 304) and Murdoch himself has a long history with media ownership, unlike the industry outsiders who are of interest for this study.
3 This archive can be accessed at http://pioneerpress.newsbank.com/.
4 Although considered broad, “political news” did have demarcations in during the coding process, meaning two articles on the same event could be coded differently depending on the focus of each article. For example: The killing of the brother of Afghanistan’s president in 2011 was reported in different ways at different newspapers. The Post’s article focused on the deceased’s role as a behind-the-scenes figure in the country’s governance; this article was coded as “politics”. The Times’ article from the same day, meanwhile, focused on the assassination itself; this article was coded as “not politics”.
5 A sixth ratio measurement – the amount of speculation – was developed for this paper, based on Barnhurst and Mutz (1997). Their method asked the amount of news analysis to be rated on a scale from 1, or “highly specific, event-centered coverage”, to 10, or “very general news analysis”. (1997: p. 30). Although more nuanced than the high-medium-low rating specified by Salgado and Strömbäck, it, too, was considered highly subjective compared to a ratio-type measurement.
Initially, the amount of speculation was operationalized in this paper’s codebook as a simple count of sentences that used “modal auxiliary verbs (will, would, can, could, shall, should, may, might) or other phrases that signal
prediction (if/then, seems, appears, likely) not based on fact” (See Appendix). However, during coding, this indicator proved unwieldy, and the result was excluded from analysis.
References
Archer, A.M. & Clinton, J. (2017). Changing Owners, Changing Content. Political
Communication, 35(3), 1-18. DOI: 10.1080/10584609.2017.1375581.
Bagdikian, B.H. (2000). Preface to the Sixth Edition. In The Media Monopoly (6th ed.). Boston, MA: Beacon Press.
Bailard, C.S. (2016) Corporate Ownership and News Bias Revisited: Newspaper Coverage of the Supreme Court’s Citizens United Ruling. Political Communication, 33(4), 583–604. DOI: 10.1080/10584609.2016.1142489.
Barnhurst, K.G., & Muntz, D. (1997). American Journalism and the Decline in Event-Centered Reporting. Journal of Communication. 47(4), 27-53. DOI:
10.1111/j.1460-2466.1997.tb02724.x.
Beaujon, A. (2012). Larry Kramer is USA Today’s new president and publisher. Poynter. Retrieved from [https://www.poynter.org/reporting-editing/2012/larry-kramer-is-usa-todays-new-president-and-publisher/].
Blumler, J.G., and Gurevitch, M. (1975). Towards a comparative framework for political communication research. In Political Communication. Issues and Strategies for
Research. S.H. Chaffee (Ed.) Beverly Hills, CA: Sage, 165–193.
de Vreese, C., Esser, F., & Hopmann, D.N. (2017). Preface to Comparing Political Journalism, xiii-xiv. New York, NY: Routledge.
Djerf-Pierre, M., and Weibull, L. (2008). From public educator to interpreting ombudsman. In: Strömbäck, J., Orsten, M., and Aalberg, T. (Eds.). Communicating Politics. Political
Communication in the Nordic Countries. Gothenburg: Nordicom, 195–214.
Dunaway, J. (2008). Markets, Ownership, and the Quality of Campaign News Coverage. The
Journal of Politics, 70(4), 1193-1202. DOI: 10.1017/s0022381608081140.
Dunaway, J. (2013). Media Ownership and Story Tone in Campaign News. American Politics
Research, 41(1), 24-53. DOI: 10.1177/1532673X12454564.
Exploring the Amazon.. (2015). The Economist. Retrieved from
[https://www.economist.com/business/2015/05/28/exploring-the-amazon].
Gosling, E. (2012). Wolff Olins creates new USA Today branding. Design Week. Retrieved from [https://www.designweek.co.uk/issues/may-2012/wolff-olins-creates-new-usa-today-branding/].
Hallin, D.C. & Mancini, P. (2012). Comparing Media Systems: A Response to Critics. In
Handbook of Comparative Communication Research. F. Esser & T Hanzitsch (Eds.).
Hanretty, C. (2014). Media outlets and their moguls. European Journal of Communication, 29(3), 335-350. DOI: 10.1080/17512786.2014.882063.
Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data. Communication Methods and Measures, 1, 77-89.
Herman, E.S., & Chomsky, N. (2002). Manufacturing Consent: The Political Economy of the
Mass Media [PDF file]. New York, NY: Pantheon Books.
Kennedy, D. (2018). The Return of the Moguls: How Jeff Bezos and John Henry Are Remaking
Newspapers for the Twenty-First Century. Lebanon, NH: University Press of New
England. Retrieved from [https://www.worldcat.org/title/return-of-the-moguls-how-jeff-
bezos-and-john-henry-are-remaking-newspapers-for-the-twenty-first-century/oclc/1004957482/viewport].
McKnight, D. (2010). Rupert Murdoch’s News Corporation. Historical Journal of Film, Radio
and Television, 30(3), 303-316. DOI: 10.1080/01439685.2010.505021.
McManus, J. (1995). A Market-Based Model of News Production. Communication Theory, 5(4), pp. 301-338. DOI: 10.1111/j.1468-2885.1995.tb00113.x.
Patterson, T.E. (1993) Out of Order: An incisive and boldly original critique of the news media's
domination of America's political process. New York: Knopf.
Picard, R.G., & van Weezel, A. (2008). Capital and Control: Consequences of Different Forms of Newspaper Ownership. International Journal on Media Management, 10(1), 22–31. DOI: 10.1080/14241270701820473.
Paid Circulation of Daily Newspapers in the U.S. 2016. (2018). Statista by Alliance of Audited Media Retrieved from [https://www.statista.com/statistics/183422/paid-circulation-of-us-daily-newspapers-since-1975/].
Reese, S.D. & Shoemaker, P.J. (2016). A Media Sociology for the Networked Public Sphere.
Mass Communication and Society, 19(3), 389-410. DOI: 10.1080/15205436.2016.
1174268.
Rohal, M. (2013). PEJ Renamed Pew Research Center’s Journalism Project. Pew Research
Center. Retrieved from
[https://www.pewresearch.org/2013/10/03/pej-renamed-as-pew-research-centers-journalism-project/].
Salgado, S., & Strömbäck, J. (2011). Interpretive journalism: A review of concepts, operationalizations and key findings. Journalism. 13(2), 144-161. DOI: 10.1177/1464884911427797.
Salgado, S., Strömbäck, J., Aalberg, T., & Esser, F. (2017). Interpretive Journalism. In
Snider, M. (2016). Former USA TODAY publisher named TheStreet's interim CEO. USA Today. Retrieved from [https://eu.usatoday.com/story/money/2016/02/23/former-usa-today-publisher-named-thestreets-interim-ceo/80795524/].
Strömbäck, J. (2008). Four Phases of Mediatization: An Analysis of the Mediatization of Politics. Press/Politics, 13(3): 228-246. DOI: 10.1177/1940161208319097.
Wagner, M.W. & Collins, T.P. (2014). Does Ownership Matter? Journalism Practice, 8(6), 758-771. DOI: 10.1080/17512786.2014.882063.
Zaller, J. (2003) A New Standard of News Quality: Burglar Alarms for the Monitorial Citizen.
Appendix A – Codebook
0 Coder ID - Please enter the coder ID assigned to you.
1 What is the article ID number? Use the name of the file.
2 Before proceeding, read through the entire article once.
Do not edit or alter the article at this point, unless for ease of readability (such as changing the font size or correcting document formatting issues).
Do not at any point save the file with any changes you made to the file.
When you have finished coding the article, you will close the file without saving.
3 Is this article about politics?
Politics includes anything related to the functions of government and the discussion of policy at the local, state, national and international levels (and every step in between), including but not limited to elections and referenda, citizens' initiatives, the actions of political parties and
individual politicians, judicial interpretations of statutes at the U.S. or state Supreme Court levels (not jury trials or criminal court rulings), and international diplomacy
o Yes (1)
o No (2)
If No (2), Go To: Q3a If Yes (1), Skip To: Q4
3a You have finished coding this article.
Before clicking the continue button, close the Microsoft Word of the article you just coded file without saving -- i.e., click "Don't save".
Upon you click the continue button, you will be automatically redirected to the start of the survey to code the next article. There is no need to close this Internet browser tab.
4 Does the article have an analytical lead?
The lead is the first paragraph(s) of the article. An analytical lead contains an opinionated explanation (the why) of news events, rather than a description (the who, what, where, where). Example of an analytical lead (coded as "Yes"):
There is a giant gap between what many of the world's governments have promised and what they can afford. Now, the headlines from across the United States and overseas show what happens when the clunky machinery of democracy goes about trying to close that gap.
Example of a non-analytical "hard" news lead (coded as "No"):
Secretary of State John Kerry expressed optimism Sunday about reaching a historic nuclear agreement with Iran, which could be announced as soon as Monday.
Example of a non-analytical "soft" / anecdotal news lead (coded as "No"):
For five years, Gov. Nikki Haley, South Carolina's first minority governor, dismissed calls to remove the Confederate battle flag from the Statehouse lawn as a divisive issue far from her agenda.
o Yes - The lead is analytical (1) o No - The lead is NOT analytical (2)
5 Does the article include unsourced and unverifiable journalistic explanations or
interpretations of "why" behind the events or actions in the article?
An easy way to do this is to turn the headline into a "why" question, and then pinpoint a sentence in the article that answers that question. Then ask if that explanatory sentence provides an
explanation without verifiable facts or attribution to a news source.
If the journalist / author does NOT provide attribution to the explanation, code as "Yes".
Example of unsourced, unverifiable journalistic explanation (coded as "Yes"): Headline: "Budget jam closes Minn. government"
Why did a budget jam close Minnesota's government?
Pinpointed sentence: "The challenge for policymakers against that backdrop is to reach agreement on that core question of how taxes and spending will be changed without causing too much collateral damage in the process." -- policy makers failed to reach an agreement without collateral damage --> unsourced and unverifiable
Example of a verifiable explanation (coded as "No"):
Headline: "Building Trump's border wall not exactly 'easy,' experts say." Why is building the border wall not so easy?
Pinpointed sentence: "Any wall-building effort would cost billions of dollars and encounter a variety of obstacles, according to experts, documents and federal officials."
-- statement is attributed to experts, documents and federal officials
o Yes, because journalist does NOT attribute the explanations (1) o No, because journalist DOES attribute the explanations (2)
6 Does the journalist include overt unsourced commentary when covering events and actions?
Commentary includes the use of value-laden terms, subjective descriptions, or normative judgments that carry non-neutral meanings. Unsourced commentary are value-laden terms that appear to come from the journalist him/herself. Examples would be the journalist writing that something is good or harmful or false, or some other word or phrase that is not based on verifiable fact or is not sourced / quoted.
Example of unsourced value-laden commentary: It will be nearly impossible to maintain both.
The crackdown has generated a good deal of absurdity. The arrests have had a powerful chilling effect. Example of sourced value-laden commentary:
Foreign Minister Laurent Fabius expressed similar optimism about an imminent agreement. o Yes - This article includes overt unsourced commentary (1)
o No - This article does NOT include unsourced commentary (2)
7 What is the word count of the body of the article? Use the word count feature under the
Review tab in Microsoft Word. To do so, highlight the entire text of the body of the article -- exclude headlines, bylines, contributor lines or other text not part of the news article. Input the number here:
8 What is the share of the article taken by the journalists' own voice?
In the Word file, delete all sentences that contain direct and indirect quotes, sourced material and other factually verifiable information in the article; information that can be deleted includes the description information included with quotes, such as details of who is speaking or sentences that give the facts of a news event. Then, use the word count tool to count the remaining number of words. Input that number here:
9 Does the journalist include speculation about future consequences of the news events?
Speculation means unsourced and unverifiable information from the journalist that uses modal auxiliary verbs (will, would, can, could, shall, should, may, might) or other phrases that signal prediction (if/then, seems, appears, likely) not based on fact.
o Yes - The article includes journalistic speculation (1)
10 How many times does the journalist use speculation about future consequences?
Please count the number of modal auxiliary verbs (will, would, can, could, shall, should, may, might) or other phrases that signal prediction (if/then, seems, appears).
Use digits; if none, please enter 0.
Do not count verifiable / sourced prediction of the future based on fact, or speculations made in the past on events that later happened:
Republicans called on the governor to convene a special session so that a stopgap measure could be passed and a shutdown avoided. He refused. -- (This is an example of a speculation made in the past of an event that later happened.)
State parks will not open on what is normally the busy July 4 weekend.
Last week, a judge ruled that only core government functions … would continue if the government were shuttered. Congress would have 60 days to scrutinize a deal.
Do count unverifiable / unsourced speculation on a person's thoughts or future status:
In the wake of the church shootings, she seems ready to discuss racial reconciliation in the state. And it appears likely that he will be onstage in the presidential debates that begin next month.
You have finished coding this article.
Before clicking the continue button, close the Microsoft Word of the article you just coded file without saving -- i.e., click "Don't save".
Upon you click the continue button, you will be automatically redirected to the start of the survey to code the next article. There is no need to close this Internet browser tab.
Good work!
