Does the Ideology of the Newsroom Affect the Provision of Media Slant?

ABSTRACT

Although research on the provision of ideologically slanted news has focused on consumers’ demands or news ownership’s profit margins and political agendas, little focus has been paid to those individuals who create the news content: the political journalists. We use a new measure of newspaper ideology derived from a large scale survey of journalists to estimate the ideology of almost 700 newsrooms, a substantial increase over previous efforts. By estimating newsroom ideology independent of content we show that newsroom ideology influences the responsiveness of newspapers to the demands of readers. We find that newsroom ideology has an effect on the ideological slant of news content even after controlling for consumer preferences. While consumer demand influences the ideological content of the news, the ideology of the newsroom that produces the news skews the responsiveness to the demands of readership and ultimately affects the production of ideological slant in the news.

KEYWORDS:

• Media Slant
• Journalism
• Latent Variable Modeling
• Differential Item Functioning
• Newspaper Ideology

Acknowledgments

We are grateful for the research assistance of Michael Christiansen, Jessica Khan, Aubrey Kohl, and Zach Stoll. A previous version of this paper was presented at the Midwest Political Science Association Conference in 2019. We are also thankful for comments and insights from Rob Bond, Josh Darr, Yphtach Llekes, Josh McCrain, and Dominik Stecula. All errors, of course, remain our own.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Data Availability Statement

The data described in this article are openly available at https://doi.org/10.7910/DVN/LQMGBG.

Open Scholarship

This article has earned the Center for Open Science badges for Open Data, Open Materials and Preregistered. The data and materials are openly accessible at https://doi.org/10.7910/DVN/LQMGBG.

Notes

1. See (Puglisi & Snyder, 2015) for a thorough overview of the debate about the influence of supply and demand factors in the provision of ideological news content.

2. Most closely related to our work here on newsroom ideological views, Baron’s (2006) work uses a formal model to consider that ideological bias might come from journalists and to show that media ownership would be willing to allow journalists leeway in what to write in exchange for lower salaries as a way of maximizing profits.

3. For context, (Groseclose & Milyo, 2005) create ideological measurements for 20 media sources, (Budak et al., 2016) for 15 media sources, (Larcinese et al., 2011) for 140 media sources, and (Gentzkow & Shapiro, 2010) for 433 media sources. In contrast, our process resulted in the ideological placement of 695 newspapers, 262 (or 38%) more than the largest previous estimate.

4. There are also other measures of newspaper ideology that are derived explicitly from news content. We do not consider these measures here because their nature does not allow for the differentiation between content and ideology and we are interested in the effects of ideology on content.

5. While we do not know exactly the percentage of papers in the sample of 695 newspapers that do not make national political endorsements, we suspect it is high given that about 47% of respondents in our survey indicated that their newspaper rarely or never covers statewide or congressional campaigns.

6. The survey resulted in a sample that mirrors previous studies of journalists in many ways. Consistent with previous surveys of journalists, we find that a plurality of journalists (46%) indicate that they are ideological moderates and that liberals (43%) outnumber conservatives (11%) in the sample (Willnat & Weaver, 2014).

7. The placement of newspapers on an ideological scale was at the end of the survey which was about 10 minutes long. The questions immediately preceding it were about a journalist’s own partisanship and ideology, professional ambitions, and the frequency of dealing with political news in the respondent’s current position. The other questions on the survey addressed how the journalist would cover particular stories and how newsworthy different types of political stories would be to editors and readers. The focus of the survey on the newsmaking process rather than on news content suggests that journalists would have the newsmaking process prominent in their mind. The full survey questionnaire is available in the online appendix.

8. Placement of a newspaper might also be affected by norms of objectivity. However, for this to affect correct placement, it would also have to have no effect on the placement of other similar outlets. Moreover, as noted further on consistent with our modeling assumptions, as a reporter’s self-reported ideology moves from liberal to conservative, so too does the overall baseline for how conservative or liberal a paper is.

9. When looking at the Washington Post and Wall Street Journal we incorporate the error in the measure for them, as this reflects that there is variation in how these two papers are scored. We do not do the same when examining the median paper as the score of the Snoqualmie Valley Record is measured with a sizable amount of error and so incorporating it would create the appearance that almost 95% of papers are indistinguishable from the median.

10. In the appendix we replicate these results using Designated Market Area vote share. Those results provide an even stronger effect of newsroom ideology on news content than those reported here.

11. We follow (Martin & McCrain, 2019) and use bigrams (two word terms) and an elastic net regression to predict ideology.

12. We drop days with less than 10 shared terms between the Congressional Record and the newspaper content.

13. These functions can be written as a set of additive parameters and so can be easily estimated.

14. See Footnote 3 for comparisons.

Additional information

Notes on contributors

Hans J. G. Hassell

Hans J.G. Hassellis an Associate Professor of Political Science and the Director of the Institute of Politics at Florida State University.

Matthew R. Miles

Matthew R. Miles is an Associate Professor of Political Science at Brigham Young University - Idaho.

Kevin Reuning

Kevin Reuning is an Assistant Professor of Political Science at Miami University of Ohio.