Press Secretaries, Journalists, and Editors: Shaping Local Congressional News Coverage

Abstract

The news media plays a key role in American democracy, often serving as the primary means by which voters learn about their elected representatives. However, the news media varies in its coverage of representatives, presenting voters with more frequent and favorable information about some House members than others, which may in turn influence voters' decisions at the polls. Although many scholars have examined the determinants of congressional news coverage, few have focused on the role of the actors who perhaps exert the most direct effect on such coverage: congressional press secretaries, journalists, and editors. In this study, I explore the influence of these actors on the tone and frequency of local congressional news coverage. I rely on data from two sources: (a) a content analysis of newspaper coverage of 100 representatives during the month prior to the 2006 election and (b) in-depth interviews with 51 congressional press secretaries and 22 journalists. These sets of data illustrate the important roles of both newspaper staff and congressional press secretaries in shaping the coverage House members receive. I conclude by discussing the implications of the findings for U.S. representatives and their constituents.

Keywords:

• U.S. representatives
• news media
• editors
• journalists
• congressional press secretaries

Acknowledgments

The author would like to thank Kim Fridkin, Pat Kenney, Jason Barabas, and the anonymous reviewers for their helpful comments; Amelia White, Jill Carle, and Elizabeth O'Callaghan for their research assistance; and Julie Schlafer for her editing services.

Notes

1. Vinson (2003) is a notable exception; however, her analysis suggests (contrary to the results presented here) that representatives' efforts to shape the coverage they receive have limited impact, particularly on the orientation of local coverage. Differences found here and in Vinson's piece may be due to a number of factors, including variation in the measures of “effort,” the examination of different media outlets, and the time period of analysis.

2. The representatives included in this study were selected for a broader project that required significant numbers of women and racial and ethnic minorities. Members of the U.S. House were stratified by gender, race, and ethnicity; within these strata, representatives were randomly selected. As such, this sample includes 22 Latinos and 28 African Americans. Largely due to the oversample of women and minorities, a significant portion (69) of the representatives in this sample are Democrats. To correct for this oversample, the multivariate analyses presented in this article are weighted by the party attachment, race, ethnicity, and gender of the representatives.

3. The vast majority of interviews were conducted prior to the midterm election; however, a small number (less than 5) of the press secretaries would not agree to be interviewed until after the election.

4. In an effort to ensure that respondents were as forthcoming as possible, all interviews were conducted on the condition of anonymity. As a result, the identities of the press secretaries (as well as the specific representatives they worked for) and the journalists interviewed are withheld.

5. Press releases were downloaded from representatives' Web sites during the period of analysis.

6. The correlation was .443 (p < .05, n = 42).

7. The Pew Research Center (2006) measured President Bush's job approval rating at 37% in October of 2006.

8. For example, an NBC News/Wall Street Journal poll of 1,006 registered voters in October of 2006 indicated a congressional approval rating of approximately 16% (http://online.wsj.com/public/resources/documents/nbc-wsj_electpoll101806A.pdf).

9. See “Newspaper Publisher's Statements: Circulation Records for the Six Months Ending September 30, 2006” (http://www.accessabc.com/).

10. No representatives in the subsample were referred to the Ethics Committee during the period of analysis; as a result, this variable was removed from all subsample analyses.

11. To examine whether the impact of press releases (and editorial endorsements) diminished in the subsample due to the decreased sample size or the inclusion of a variable measuring representative-reporter relationships, all of the models presented were reestimated using the smaller sample (without the inclusion of reporter-representative relationships). In these reestimated models, the impact of the press release and editorial endorsement variables was no longer statistically significant, indicating that the limited sample (not the inclusion of representative-reporter relationships) is responsible for these changes.