This study expands the theory of second-level agenda setting to include affective framing of candidates conveyed through visual information during the 2000 presidential campaign. Network TV news coverage included nonverbal behavior for Al Gore that was more positive than George Bush's, and those who watched more were significantly more likely to hold attitudes that mirrored the media portrayals.
In the era of television politics, there is a growing trend of style over substance, personality over issues, and emotion over information. Images candidates project and how news media portray them receive attention for the effect they may have. Second-level agenda-setting theory recognizes the importance of affective perceptions, emphasizing these attributes and their tone as being just as important as cognitive-level issue salience. Whereas first-level agenda setting suggests a role for media in deciding what issues the public is aware of, with researchers focusing on amount of coverage, second-level agenda setting suggests the media also frame attributes of these issues, thus affecting how the issue is defined.1
A growing body of research has shown that the affective attributes reported in the media about newsmakers such as political candidates influence the attributes the public associates with those newsmakers.2 In these second-level studies, affective attributes of candidates have been examined in terms of one mode of communication-words. Even when the primarily visual medium of television is studied, affective attributes of candidates are examined in the verbal content only. This is a notable gap, given that nonverbal communication is especially adept at communicating affective information.3
This study, then, proposes an expansion of second-level agenda-setting theory by examining affective framing of candidates through visual rather than verbal information.4 It pairs content analysis of the 2000 presidential candidates' nonverbal behavior with survey data from the National Election Studies, correlating the media affective agenda communicated through visuals with audience affective impressions of candidates. The research expands the study of agenda setting's affective dimension to a method of communication shown to clearly and convincingly convey affect-the visual mode.5 As Baran and Davis wrote, "Stories are often complex combinations of visual and verbal content-all too often the visual information is so powerful that it overwhelms the verbal."6
We test two ideas-first, that television news images of candidates exhibiting positive or negative behaviors such as gestures and facial expressions convey impressions that might be flattering or unflattering.7 Second, we test whether the second-level agenda conveyed visually corresponds with the public's affective impressions of the candidates. We do not propose that the visual is more important than the verbal, but that visual information can serve as additional, important information that will be considered as audiences evaluate messages. We propose that both channels of communication work together to contribute to agenda-setting effects, and that understanding of the process is possible only by studying both modes of communication.
This study has a practical implication because late-deciders may be more influenced by visual displays.8 Visual presentations may have played an unusually important role in a race too close to call for most of the campaign, and may help us better understand future political races where no clear frontrunner is evident.
Theory. For more than thirty years, scholars have studied agenda-setting effects of news on public opinion. Agenda-setting research examines how the media set the agenda of public opinion, a process McCombs and Shaw9 called the transfer of issue salience. Since then, hundreds of studies have explored this link. At first only public issues were studied; more recently the focus of agenda-setting studies has been "objects," a rubric encompassing other subjects. …