To determine the public agenda in Chapel Hill during the 1968 presidential election, a survey was conducted among a sample of randomly selected undecided voters. Only undecided voters were interviewed because this new agenda-setting hypothesis went against the prevailing view of media effects. If this test in Chapel Hill failed to find agenda-setting effects under rather optimum conditions, voters who had not yet decided how to cast their presidential vote, there would be little reason to pursue the matter among the general public, where long-standing psychological identification with a political party and the process of selective perception often blunted the effects of mass communication during election campaigns.
In the survey, these undecided voters were asked to name the key issues of the day as they saw matters, regardless of what the candidates might be saying. The issues named in the survey were ranked according to the percentage of voters naming each one to yield a description of the public agenda. Note that this rank ordering of the issues is considerably more precise than simply grouping sets of issues into those receiving high, moderate, or low attention among the public.
The nine major news sources used by these voters were also content analysed. This included five local and national newspapers, two television networks and two news magazines. The rank order of issues on the media agenda was determined by the number of news stories devoted to each issue in recent weeks. Although this was not the very first time that survey research had been combined with content analysis to assess the effects of specific media content, their tandem use to measure the effects of mass communication was rare at that time.
Five issues dominated the media and public agendas during the 1968 US presidential campaign – foreign policy, law and order, economics, public welfare, and civil rights. There was a near-perfect correspondence between the rankings of these issues by the Chapel Hill voters, and their rankings based on their play in the news media during the previous twenty-five days. The salience of five key campaign issues among these undecided voters was virtually identical to the salience of these issues in the news coverage of recent weeks.
Moreover, the idea of powerful media effects expressed in the concept of agenda setting was a better explanation for the salience of issues on the public agenda than was the concept of selectivity, which is a keystone in the idea of limited media effects. To be clear, agenda setting is not a return to a ‘magic bullet’ or ‘hypodermic needle’ theory of all-powerful media effects. Nor are members of the public regarded as automatons waiting to be programmed by the news media. But agenda setting does assign a central role to the news media in initiating items for the public agenda. Or, to paraphrase Lippmann, the information provided by the news media plays a key role in the construction of our pictures of reality. And, moreover, it is the total set of information provided by the news media that influences these pictures.
In contrast, the concept of selectivity locates the central influence within the individual and stratifies media content according to its compatibility with an individual’s pre-existing attitudes and opinions. From this perspective, it is often assumed that the news media do little to alter the issue priorities of individuals because individuals maximize their exposure to supportive information and seek out news about issues that they already deem important. For instance, during an election, voters are expected to pay the most attention to those issues emphasized by their preferred political party.
Which does the public agenda more closely reflect? The total agenda of issues in the news, which is the outcome hypothesized by agenda-setting theory? Or the agenda of issues advanced by a voter’s preferred party, which is the outcome hypothesized by the theory of selective perception?
To answer these questions, those undecided Chapel Hill voters who had a preference (albeit not yet a firm commitment to vote for a candidate) were separated into three groups – Democrats, Republicans, and supporters of George Wallace, a third-party candidate in that election. For each of these three groups of voters, a pair of comparisons was made with the news coverage on the CBS television network: the issue agenda of that voter group compared with all the news coverage on CBS, and the issue agenda of the group compared with only the news on CBS originating with the group’s preferred party and candidate. These pairs of comparisons for CBS were repeated for NBC, the New York Times, and a local daily newspaper. In sum, there were a dozen pairs of correlations to compare: three groups of voters times four news media.
Which was the stronger correlation in each pair? The agenda-setting correlation comparing voters with all the news coverage, or the selective perception correlation comparing voters with only the news of their preferred party and candidate? Eight of the twelve comparisons favoured the agenda-setting hypothesis. There was no difference in one case, and only three comparisons favoured the selective perception hypothesis. A new perspective on powerful media effects had established a foothold.
The accumulated evidence
A year after publication of the Chapel Hill study, Theodore White’s last instalment of The Making of the President series details the agenda-setting influence of the news media in the excerpt shown in Box 1.2. Since the modest beginnings in Chapel Hill during the 1968 presidential election, more than 500 separate scientific works on agenda setting have accumulated, spanning all six continents, including political and non-political settings, across a variety of media, and dozens of issues.11 In 2011, the 75th anniversary celebration of Public Opinion Quarterly noted that the Chapel Hill study was the most cited article ever published in the journal.12 By March 2020, this article alone had over 12,000 citations in Google Scholar. Some authors even speak of the ‘agenda-setting juggernaut’13 to highlight the popularity of the agenda-setting concept when studying the influence of journalists and the news they produce. And it may well be the only theory in journalism studies to have a scholarly journal entirely devoted to it, The Agenda-Setting Journal.
Despite reams of published research, the basic idea of the theory has remained straightforward: Agenda setting refers to the process by which the elements (e.g. issues, public figures, companies, or government institutions) that are deemed relevant by the news media as well as the attributes used to describe these elements often become relevant to public opinion, too. Importantly, the basic agenda-setting hypothesis has been widely documented. The latest meta-analysis – a technique that pools in a statistically meaningful way the results of separate studies – found an average correlation of +0.49 between the media and public agendas.14 To put this number in perspective, consider that the mean effect estimate for human communication phenomena is +0.21.15 This robust collection of evidence also documents the time-order and causal links between the media and public agendas in finer detail. Here is a sampling of that evidence.
The 1972 US presidential election in Charlotte
To extend the evidence for agenda setting beyond the narrow focus on undecided voters in Chapel Hill and their media sources during the autumn 1968 election, a representative sample of all voters in Charlotte, North Carolina, and their news media were examined three times during the summer and autumn of 1972.16 Two distinct phases of election-year agenda setting were identified. During the summer and early autumn, the daily newspaper was the prime mover. With its greater capacity – scores of pages compared to half an hour for network television news – the Charlotte Observer influenced the public agenda during the early months. Television news did not. But in the final month of the campaign, there was little evidence of agenda setting by either the local newspaper or the television networks.
In addition to documenting the agenda-setting influence of the local newspaper on the public, these observations across the summer and autumn of that election
