Election Study data, Nie, Verba, and Petrocik, found that citizens' references to the personal characteristics of candidates were the most frequent type of response and had been consistently high in election after election.29
Schema theory (which we discussed in Chapter 2) certainly expanded the social-psychological approach to more fully analyze the way people mentally process political communications and make sense of the political world. But the Democracy '92 team thought, in general, a richer—yet realistic—conception was needed of the citizen's role in elections and in the reception of and response to political communications. The team focused on the construction of candidate images in peoples' minds—images in a rather inclusive sense. They recognized the reality of peoples' emphasis on the more accessible information on and visual and aural impressions of the personal nature of candidates. But they also recognized that people tend to incorporate a range of "considerations" in their candidate evaluations, including some political party and issue content and assessments of past performance, as Popkin has suggested.30 They also recog¬nized, as have scholars from Graber to Lodge, McGraw, and Stroh,31 that peo¬ple do not tend to remember much in the way of details about what they were exposed to regarding a candidate. Rather, people tend to pay varying attention to the communications to which they happen to be exposed and gradually con¬struct a summary judgment of each candidate, and it is that summary judg¬ment and only the most important basic considerations that they specifically remember.
The study sought to analyze the discourses of the campaign and what role each type and source of discourse plays in the construction of candidate images. That is, candidates seek to construct messages and visual and aural impressions that build the type of image the candidate wants the voters to have of that can¬didate and of his or her opponent(s); these messages and impressions are com¬municated through the candidate's own communications—in ads, debates, speeches, and so on. Thus, it is important to content analyze the most signifi¬cant of those candidate presentations. But the candidate's image is also shaped by how the candidate appears in the news media, communications which are mediated by news personnel, thus, news media stories, editorials, cartoons, and so on also must be content analyzed. Further, what the impact of those commu¬nications on the public is must be analyzed. But the public is not simply a sponge, and takes an active part in the process. Thus, the public's information and opinions on the electoral choices must be looked at, and how people con¬struct their images of the candidates requires in-depth monitoring and analysis. Before we discuss the specific methods used, we should note that the team also recognized that the news media, the candidates, and the voters each get feed¬back from the others, which has an impact on the way they operate and commu¬nicate, so that factor is important to consider as well.
279
Chapter 10 The Media in Elections I
Earlier, we noted some weaknesses in various methods of discerning the public's response to electoral communications. Is there a way to overcome those weaknesses? The answer by the Democracy '92 study was a multimethod approach. On the news media side, the study employed monitoring and content analysis similar to the Patterson approach, coding news stories, editorials, edito¬rial cartoons, and so on for media source, date, placement, size of story, who and what issues are mentioned, etc. More detail was added regarding various aspects of the visual dimension of the news, including coding for tone of visuals, as well as overall tone of the story for each candidate. More detail was also added regarding aspects of the leadership, background, and personal characteristics of the candidates; what the "main focus" of the story was; number and length of candidate sound bites; and so on. Recognizing the national as well as local nature of news sources, the study monitored the prime evening network news shows, including CNN, from February 1 through the week of election day. The study also expanded the local areas monitored, studying the leading one or two local TV news shows and the area newspaper (two, in Boston's case) in four dif¬ferent local areas: Los Angeles; Boston; Winston-Salem, North Carolina; and Fargo, North Dakota-Moorhead, Minnesota. These were in four different regions of the country, in four quite different political subcultures, and ranged in size from one of the nation's largest metropolitan regions to a second-tier big city to two small city areas, one in prime agricultural country. Thus, a good range of areas was included. At the national level, recognizing the much greater use of new venues for candidate appearances, national TV talk, magazine, and Sunday news interview shows were also monitored and analyzed.
The multimethod approach is especially apparent in the aspect of the study that looked at the public. In each of the four local areas, three different methods were used to record and probe the public's understanding of election-related matters in general and their construction of candidate images in particular: (1) opinion surveys were conducted with a random sample of the public, (2) in-depth interviews were conducted (averaging an hour or more each) with ten to fifteen people selected to have a range of demographics and political interest levels, and (3) focus groups of twelve people each were also conducted. These instruments were conducted in waves in order to cover the significant stages of the election period. A baseline set of interviews was conducted in the third week of January. Just before the presidential primary or caucuses in each respective state, early in the general election, and late in the general election, each of the three methods described was used in each area (with a partial exception in one area in the case of the polling). Using all three methods to ascertain what people were thinking allowed the study to draw on the strengths of each method and to obtain a cross-check, to some extent, on what findings could be generalized. Thus, the opinion surveys enabled a set of focused questions to be posed to a representative sample of the local publics. The interviews, on the other hand,
280
The Media in Elections: Methods of Study
enabled a much deeper review of the way people thought about the communica¬tions and the way they constructed their images—largely in their own words and concepts, since the interviews were based on a set of rather general questions with generic probes. Those interviews relied on peoples' memories over many weeks. The focus groups, however, with their use of videotaped material from news shows, ads, and so on as direct stimulus; materials, allowed a closer look at the specific patterns of response to such communications. The focus group discussions, which were videotaped, also enabled the team to assess how people discussed electoral matters in a group setting.
Hnally, the candidates' ads, from the nomination phase as well as the gen¬eral election phase, were content analyzed using categories, scales, and so on of the same nature as was used in the news coding. Indeed, in all the instruments, the basic questions, question wording, scales used, and matters monitored were coordinated as much as possible; this afforded maximum opportunity to accu¬rately compare the different modes of communication. Perot's long-form ads or "infomercials" (30-minutes) and the Clinton and Bush convention speeches were also content analyzed, using a somewhat simpler approach. Each of those modes of candidate message and presentation were compared with the others to see the extent and manner in which the candidates sought to coordinate and reinforce their messages, or offered different messages in different modes of campaign communications, or did not have their message act together. One fur¬ther analytical dimension was added, a dimension which has had little attention paid to it by scholars (in part because of the difficulty of obtaining the data). The team gathered the data on ad buys—buys made by the campaigns of specific times for airing specific ads on network TV and on the local TV' stations in the four areas studied. This enabled the team to analyze just which ads were com¬municated to the public to what extent and what images and issues they con¬veyed—and it gave a unique insight into the strategies of the campaigns.
One problematic element in this study was the use of February 1 as the date for the beginning of the analysis of news stories (it was originally intended to be January 15, but some problems in the monitoring process required the later start date). In general, news coverage of importance for a presidential election certainly begins before February 1; in 1992, this was more true than ever, as the first intense week of Clinton scandal coverage came during the last week in Jan¬uary. Another methodological problem in this study is, in fact, common to almost all content analyses of news media in elections; it will be discussed in a subsequent section of this chapter.
The Experimental Method
The Democracy '92 study was an unprecedentediy comprehensive approach to the study of the media in elections. But even with all the different methods used
281
Chapter 10 The Media in Elections I
to obtain the most accurate data and analyses of the various communications and events and their impacts, there is still considerable difficulty in being certain that particular causes are linked to particular effects. This is because each fac¬tor—variable in formal terms—is not neatly controlled for. With careful design and operation, the experimental method can, however, control each variable, vary it systematically, and provide more precise evidence on the impact of the independent variable on the dependent