hypothesis that the public agenda influenced the newspaper agenda. When there are observations of the media agenda and the public agenda at two or more points over time, it is possible to compare simultaneously the cross-lag correlations measuring the strength of these two competing causal hypotheses. For example, the influence of the newspaper agenda at time one on the public agenda at time two can be compared with the influence of the public agenda at time one on the newspaper agenda at time two. In Charlotte, the agenda-setting hypothesis prevailed.
The agenda of issues during the 1972 presidential campaign included three very personal concerns – the economy, drugs, and bussing to achieve racial integration of the public schools – and four issues that were more remote – the Watergate scandal, US relations with Russia and China, the environment, and Vietnam. The salience of all seven issues among the public was influenced by the pattern of news coverage in the local newspaper.
The 1976 US presidential election in three communities
An intensive look at an entire presidential election year followed in 1976 and again highlighted variations in the agenda-setting influence of the news media during different seasons of the year.17 To capture these variations, panels of voters were interviewed nine times from February through December in three very different settings: Lebanon, New Hampshire, a small town in the state where the first presidential primary to select the Democrat and Republican candidates for president is held each election year; Indianapolis, Indiana, a typical midsized American city; and Evanston, Illinois, a largely upscale suburb of Chicago. Simultaneously, the election coverage of the three national television networks and the local newspapers in these three sites was content analysed.
In all three communities the agenda-setting influence of both television and newspapers was greatest during the spring primaries, when voters were just beginning to attend to the presidential campaign. A declining trend of media influence on the public agenda during the remainder of the year was particularly clear for the salience of seven relatively remote issues – foreign affairs, government credibility, crime, social problems, environment and energy, government spending and size, and race relations. The salience of more personal matters, such as economic issues, remained high for voters throughout the campaign, regardless of their treatment by newspapers and television. Personal experience can be a more powerful teacher than the news media when issues have a direct impact on people’s lives.
Although these detailed examinations of the issues on the public agenda help us understand the variations in the agenda-setting influence of the news media, the specific issues change from election to election. So it is useful to have a summary statistic that will allow us to compare the degree of agenda setting taking place in different settings. One of the most common measures used by scholars exploring the agenda-setting role of the news media is the correlation statistic. This statistic summarizes precisely the degree of correspondence between the ranking of issues on the media agenda – which issue received the most news coverage, which issue the second most coverage, etc. – and the ranking of those same issues on the public agenda – which issue most members of the public regard as most important, which issue ranks second among the public, etc.
The possible range of scores for the correlation statistic is from +1.0 (perfect correspondence) through 0 (no relationship at all) to −1.0 (a perfectly inverse relationship). Agenda-setting theory predicts a high positive correlation between the media agenda and the subsequent public agenda.
Using this correlation statistic to summarize a key finding from the intensive year-long look at the 1976 presidential election in three different communities, we see that, during the spring primaries when the agenda-setting influence of both television and newspapers was at its peak, the correlation between the national television agenda and the subsequent voter agenda was +0.63. That is a significant degree of influence. In contrast, the correlation between the agendas of the three local newspapers read by these voters and the voters’ agenda of public issues was only +0.34. Nevertheless, this was the peak period for the newspapers. Although it is fashionable to attribute great influence to television in many aspects of life, do not rush to generalize this particular finding about the relative influence of television and newspapers. The final section of this chapter will present a more cautious and comprehensive picture of the agenda-setting role played by various communication media.
These extensive observations of the 1976 presidential campaign across the entire election year provide another opportunity to compare the core hypothesis of agenda-setting theory that the media agenda influences the public agenda with the competing causal hypothesis that the public agenda influences the media agenda. In contrast to the agenda-setting correlation of +0.63 noted above for national television during the spring, over the same time period the correlation between the public agenda and the subsequent national television agenda is only +0.22. The difference between the two is further amplified by comparison with the Rozelle–Campbell baseline, a statistic indicating the value to be expected by chance alone. In this instance the Rozelle–Campbell baseline is +0.37. The agenda-setting correlation is far above this baseline. Its rival is below the baseline. For newspapers, the rather low agenda-setting correlation of +0.34 nevertheless compares quite favourably with the rival correlation of +0.08. The Rozelle–Campbell baseline in this instance also is +0.08. In both of these instances, the evidence corroborates the causal influence of newspaper and television issue agendas on the public agenda.
These initial empirical efforts to map the agenda-setting role of the mass media encompassed three consecutive US presidential campaigns. Election settings were not selected because of any assumption that agenda-setting effects are limited to elections, but rather because national elections create a natural laboratory for the examination of media effects. During a national election there is a continuing massive barrage of messages on public issues and other aspects of politics. If these messages are to have any significant social effects, the effects must occur by election day.
In addition to these advantages for studying media effects, there is also an enduring tradition of scholarship on the role of mass communication in national elections that began with the seminal studies of Lazarsfeld and his colleagues, first in Erie County during the 1940 US presidential election and then in Elmira, New York, during the 1948 US presidential election. For all these reasons, the initial examinations of agenda setting were conducted in election settings.
However, the agenda-setting role of the news media is limited neither to elections nor to the United States, nor even to the arena of political communication broadly defined. American presidential elections were just the starting point. The phenomenon of agenda setting, a continuous and inadvertent by-product of the communication process, is found in both election and non-election settings, at both the national and local levels, in a wide array of geographical settings worldwide, and increasingly for a broad array of agendas extending beyond political communication. However, for now we will focus on issue agendas, the best mapped domain of the agenda-setting role of the communication media.
National concern about civil rights
From 1954 to 1976, a 23-year span encompassing half a dozen presidential elections and all the years in between, the salience of the civil rights issue in the United States rose and fell with great regularity in response to news coverage.18 The percentage of Americans naming civil rights as ‘the most important problem’ facing the country ranged from 0 to 52 per cent in the twenty-seven Gallup polls conducted during those three decades. When this continuously shifting salience of civil rights on the public agenda was compared with the news coverage on the front page of the New York Times for the month preceding each of the twenty-seven polls, the result was a robust correlation of +0.71. Even when the influence of news coverage in earlier months is removed, the correlation remains +0.71. This is especially compelling evidence of the media’s agenda-setting role. Also note that the salience of the civil rights issue among the public primarily reflects the preceding month of news coverage, a relatively short-term response to the media agenda. Because
