Abstract
The UK press demonstrated predictably high interest in the 2004 US election, and the general course of their coverage was similar in the discussion of the significance of the race and the reasons for the outcome, the identification of important campaign issues, and the general evolution of the campaign. In specific terms, content analysis of the coverage by daily newspapers and television news verified that the coverage is still dominated by horserace factors and demonstrates predictable partisan directions. However, compared to the 2000 campaign, the coverage in 2004 is distinctive for its increased emphasis on substantive issues and its less cynical attitude toward the candidates. However, preliminary analysis of the press websites did not suggest many innovative uses of technology in campaign coverage.
Acknowledgments
This research was conducted with the aid of a grant from EDS-LSE Innovation Research Programme, Media and Communications Theme. Thanks to Helen Marsh for research assistance.
Notes
The coding frames for the two parts of the analysis were slightly different. The basic coding unit for the first section (conventions and debates) was individual stories, and each was coded for main subject and up to three subsidiary subjects, and for overall story tone and reporter tone. Across the four categories intercoder reliability was 85 per cent. The analysis from 1 October to 5 November, conducted by Media Tenor, uses two base units of analysis: the statement and the story. Stories typically contain several statements, the latter being defined as a person/institution, topic, and source. Each statement is rated on a positive/negative scale depending upon the explicit tone of the reporter (critical, cynical, disdainful) and the contextual tone (that is, whether the statement is reported as positive/negative from the viewpoint of the candidate). As any part of the statement changes (e.g. person or topic), a new statement is coded. Thus, strict comparability for the two sections can be made only at the story level. Intercoder reliability was 77 per cent for this latter analysis (1 October to 5 November).
Analysis of press coverage in the 2000 campaign covered the period from 1 August to 5 November and sampled 20 per cent of all US election stories each month in The Guardian, the Daily Telegraph, The Times and their sister Sunday papers. A total of 282 stories were coded from a population of 1200, using the same code design as was used for the debates and conventions in 2004. It was decided to exclude August from the analysis of the 2004 campaign, and thus August was excluded from comparisons of all subjects across the campaigns. The 2004 analysis took selected periods (as explained in Note 1), but examined all stories within these, a total overall of 1349 stories. Intercoder reliability was tested by the author on 10 per cent of the sample in 2000 and 10 per cent of the convention/debates sample.