Jump to ContentJump to Main Navigation
Congress and the MediaBeyond Institutional Power$

Danielle Vinson

Print publication date: 2017

Print ISBN-13: 9780190632243

Published to British Academy Scholarship Online: March 2017

DOI: 10.1093/acprof:oso/9780190632243.001.0001

Show Summary Details

(p.203) Appendix New York Times Analysis Methodology

(p.203) Appendix New York Times Analysis Methodology

Congress and the Media

C. Danielle Vinson

Oxford University Press

The two data sets used in Chapters 2 and 3 come from the New York Times. For the full data set, I randomly selected three days that Congress was in session for each year from 1977 through 2015. For the transition-year data set, I selected January and May of each year that marked the first year of a new presidential administration (1977, 1981, 1989, 1993, and 2001) or that followed a shift in partisan control of either house of Congress (1987 and 1995) from 1977 through 2001. For both data sets, my research assistants and I searched the New York Times for the selected days and months, using the search term “Representative” or “Senator” or “Speaker” to identify any article that might include public comments of members of Congress. “Speaker” was necessary as a separate term because articles often referred to the Speaker of the House by his or her formal title rather than as Representative. That issue did not arise with Senate leaders. We eliminated articles that did not pertain to the US Congress. Of the remaining articles, we kept those that quoted or paraphrased a congressional member commenting on a public issue, policy, or event related to his or her job in Congress (as opposed to the district or a member’s private life).

In each article, the unit of analysis was the member or coordinated group of members being quoted or paraphrased. Therefore, it was possible to code several cases for one article. Members were coded separately unless the article explicitly stated they were working together, as in the case of a joint statement to the press. This probably underestimates the extent of coordinated communication, but we found no consistently reliable way to divine which members were working in concert apart from (p.204) the reporter mentioning it. We ended up with 3,713 cases in the transition-year data set and 921 in the full data set.

The Content Code

For each case, we recorded the following information:

  1. I. Date

  2. II. Information about the congressional member who was speaking

    1. A. The specific person(s) speaking and the category the person(s) fit into

      1. 1. Majority party leader

      2. 2. Minority party leader

      3. 3. Committee chair

      4. 4. Ranking minority member of the committee

      5. 5. Caucus (a formal caucus within Congress, such as the Congressional Black Caucus)

      6. 6. Coalition (an informal group that comes together just for this specific issue at this time)

      7. 7. Other majority party individual

      8. 8. Other minority party individual

      9. 9. Other (specify)

    2. B. The party the congressional member belonged to

    3. C. The house of Congress the congressional member belonged to

  3. III. Information about the content and purpose of the congressional member’s comments

    1. A. Issue: a two- or three-word description of the major issue and the category that best describes that issue

      1. 1. Budget/taxes (includes deficits)

      2. 2. Economy (includes recession, unemployment, jobs outlook, and any other references to the health of the economy)

      3. (p.205) 3. Business (regulation or oversight of industries, finance, or banking issues unrelated to the economy or a specific issue category)

      4. 4. Foreign policy (treaties, trade agreements, relations with other countries)

      5. 5. Defense (includes military issues, war, national security)

      6. 6. Health care (includes Medicare, Medicaid)

      7. 7. Social policy (welfare, Social Security, and other safety-net programs)

      8. 8. Environment (all environmental issues, including conservation, pollution, nuclear waste cleanup, etc.)

      9. 9. Energy policy

      10. 10. Crime (includes law enforcement)

      11. 11. Agriculture

      12. 12. Education

      13. 13. Civil rights (civil rights and liberties, including abortion, gay rights, criminal rights, freedom of speech and religion, etc.)

      14. 14. Government affairs (includes congressional rules or procedures, ethics investigations and government scandal, and campaign finance)

      15. 15. Appointments (judicial or executive branch nominations, confirmation hearings, etc.)

      16. 16. Party politics (party fundraising, elections for party leadership, general party strategy that does not fit one of the other issues)

      17. 17. Immigration

      18. 18. Transportation

      19. 19. Other

    2. B. Reaction: whether the effort to go public was in reaction to something else (such as an event or announcement) or was an attempt to initiate something new

    3. C. Reaction to president: whether the congressional member was reacting to something the president had said or done, (p.206) and if so, whether the member was supportive of the president or opposed to the president

    4. D. Direct contact with the press: whether it appeared that the congressional member had direct contact with the press in going public: news conference, op-ed columns, interviews with talk shows or quotes in the story that were made to the reporter rather than during floor debate or committee hearings

    5. E. Work-related activity: whether comments were made in the course of the congressional members’ regular work activities, including public hearings, filibusters, and floor debate

    6. F. Goal: the apparent goal of the congressional member; this may have been explicitly stated in the article. If not, we attempted to discern the goal based on the political context.

      1. 1. Get the member’s policy preference passed or enacted

      2. 2. Stop legislation or action from taking place

      3. 3. Other

Intercoder Reliability

In the early stages of the New York Times analysis, multiple coders analyzed the same articles in the transition-year data set to help identify potential problems with the content code. These early coding efforts led to some revisions in the content code to improve reliability. Some of these changes are detailed in the relevant chapters.

Ten percent of the articles in the full thirty-nine-year data set were analyzed by a second coder to measure intercoder reliability for the revised content code used in both New York Times data sets and the television analysis in Chapter 4. On the variables related to information about the congressional member, reliability was nearly 100 percent, with discrepancies usually resulting from accidental data entry errors. Reliability on the content of congressional members’ messages was a little lower, but no variable dropped below 80 percent agreement.