studies include Vargo et al.’s (2014) examination on Twitter issue agenda during the 2012 presidential election and Vu, Guo & McCombs’s (2014) network comparison of that year’s top news stories with the public agenda. Both studies suggested significant associations between media agenda and the public agenda, lending support to the third-level agenda-setting hypothesis. The Graph I from Guo (2013) best illustrate the differences between traditional agenda-setting approaches to network agenda- setting research.
Methodologically speaking, computers play a crucial role in big-data content analysis of subject and attribute associations. Issues and keywords can be identified through assessment of random samples of vertical and horizontal media content. Computer assists the analysis of frequency of at- tributes, centrality of attributes and then the analysis of network measures the pair of issues’ strength of association. As a whole, network analysis of centrality describes how often each issue appears over time with other issues on the media agenda. McCombs expects the availability of big data analytics will provide more empirical evidence for and enrich our under- standing of third-level agenda-setting theory in the future.
Weaver: Need for Orientation and the psychology of agenda setting
Weaver’s presentation reported years of research on need for orientation (NFO), or ―the most prominent of the contingent conditions for agenda-setting effects‖ (p.67, McCombs, 2004). In 1973, the concept of NFO was first used in agenda setting by McCombs and Weaver, defined as individuals’ need to familiarize themselves with the outside world via the use news media. However, the concept was not formally introduced until the Charlotte study, when they examined the agenda-setting effects of the news coverage of the 1972 U.S. presidential election. The study suggested NFO as a psychological explanation for individual variance of agenda- setting effects.
When first introduced, NFO was conceptualized as having two dimensions: relevance and uncertainty. Relevance refers to the perceived importance of the issue in audience’s mind, and uncertainty concerns how much people find they lack the knowledge on a particular issue. In the original Charlotte study, NFO was operationalized using a ―sequential measurement,‖ which further subdivided high relevance by uncertainty and pro- duced three levels of NFO.
It was found voters with a low NFO showed a much lower correspondence between media agenda and public agenda compared to voters with a high NFO. In the years followed, research of agenda setting around the world provided similar supportive results re- garding the moderating role of NFO on agenda-setting effects.
However, many recent studies using this measurement of NFO failed to observe substantial moderation on second-level agenda-setting effects. Thus, Matthes (2006) suggested another way to implement NFO toward three aspects of agenda-setting process: the orientation needs for issues, toward the attributes of those issues, or toward journalistic assessments. In Weaver’s presentation, these aspects were referred to as objective salience, substantive salience and affective salience. Matthes (2006) maintaine the sequential measurement method when identifying NFO. That is, when relevance is low, NFO is low regardless of level of uncertainty. Uncertainty matters only when relevance is high. The new scale was tested by Matthes (2006) and Chernov, Valenzuela and McCombs (2011) as being internally reliable and it predicted the first-level agenda setting at the level of objective salience. However, Matthes’s (2006) scale could only explain the information seeking process, but not the effects of affective attributes. Weaver drew the audience attention back to the two-dimensional NFO scale using a 2X2 typology defined by high and low relevance and uncertainty. Applying this measurement, the most recent study conducted by Camaj (2014) differentiated the ―active involvement NFO‖ (high relevance and low uncertainty) from the ―active involvement NFO‖ (low relevance and high uncertainty), and found the strongest agenda-setting
effects among the active involvement NFO people
Based on the existing NFO research, Weaver presented a dual-path in- formation-processing and media-selection model (Graph II). Starting from one’s NFO, casual media exposure or deliberate media exposure is experienced, which results in accessibility bias and active inference bias respectively. Then agenda-cueing effect is more likely to be observed among those who have a low- and moderate-passive NFO while the
agenda-reasoning effect occurs among people with moderate- to high-active NFO. The two NFO groups also have different tendency in selecting niche or mainstream media, leading to different intensity of agenda setting at different levels. In conclusion, this comprehensive model to date explains the psychological processes behind agenda setting, including type of exposure, type of medium and type of agenda-setting effect, promising many empirical and theoretical directions for future research.
Shaw: Using media to monitor civic life, find personal community and create private identity
Shaw’s presentation aimed to promote an understanding of how media serve society and how we can advise media to serve government and citizens, based on the agenda-setting research. He succinctly pointed out that, ―The media do not merely cover communities, media are communities. When we collectively left our villages for life in cities and urban life, we enter a society so diverse and complex that we must rely on many mediated media to understand it.‖
Agenda-setting theory provides a useful approach, according to Shaw, in examining mass media effect because it integrates the important parts of process and effects. On the one hand, journalists frame public life in the media and on the other, journalists take the citizens’ interests into consid- eration when selecting such frames, so the daily news agenda is also about ―melding‖ public’s interests.
In reference to today’s multimedia, the concepts of vertical vs. hori- zontal medium provide easy ways to visualize how people form their own community agendas. Vertical media refer to the traditional newspapers, radio and TV,
which reach their audience like shouting from the top of a mountain to the general public. By contrast, horizontal media are created to meet personal interests, such as magazine and social media. They reach a specific public and connect people as individuals around the world. Today the world is witnessing an increased power of horizontal media to set the agenda for the public as social media are on the rise and gradually convert- ing traditional vertical media (newspapers and TV) into horizontal media of personal interests.
For example, a replication of the 1968 Chapel Hill agenda-setting study in 2008 observed that voters mixed vertical and horizontal agendas. They obtained most of their information from vertical media and some from
horizontal media. They then integrated these two news agendas into their own. In this process, the selection of vertical vs. horizontal media hinges upon individuals’ interests in elections. If we assume a closed information system consisting of direct experience and mediated sources, a formula Agenda Community Attraction (ACA) could then be used to predict how individuals form their personalized agenda communities:
Suppose the correlation between vertical media and voters was .80, the unaccounted part attributed to horizontal media would be .20. Square these values and subtract them from 1 gives us the part (.32) controlled by citizens, which is not accounted for by either vertical or horizontal media together. In this way, the formula could be used to predict the voter issues by referencing media agenda.
Based on this formula, Shaw showed a model for the Dynamics of Agenda melding and Civic Balance (Figure 3). The left side of the model suggests a relatively stable social system when correlations of vertical media and public agenda are high. In the middle is a transition period of instability as alternative community agendas gaining power. Moving toward the right shows the trend that alternative agenda community is becoming dominant as long as the correlation falls below .50. Shaw said this hypo- thetical model is yet to be tested with agenda-setting data worldwide.
Finally, Shaw talked about the evolution of media landscape and mass media audience. As the size of traditional, vertical media audience shrinks, the agenda setting effects of vertical media may decline, but those of horizontal media may increase. In addition, different generations differ from each other in their ways of using and melding various media agendas. Shaw believes that we are living at a time of historical paradigm shift. He calls for more efforts in understanding personalized melding of media agendas, and believes such an effort can help governments and citizens understand the challenge of moving toward a world that is more easily transportable by the papyrus paper – the horizontal media messages designed for special interests.
In summary, the ―three big heads‖ presented three important and ―particularly active arenas‖ (p. 782, McCombs, Shaw & Weaver, 2014) of contemporary agenda-setting research-networking agenda setting, NFO and agenda melding. Moving toward its 50 anniversary, agenda-setting theory has become well established with each of these arenas guiding current research now and in the future. McCombs, Shaw and Weaver expect future research to expand the scope beyond the traditional focus on public affairs, but there is also a need to deepen the understanding of core agenda-setting concepts.
One advice they have for young scholars is to have a programmatic research agenda. As vivid exemplars of programmatic research, all three scholars have devoted their entire careers in developing agenda-setting concepts and areas of research, and each one of them continues to look for breakthroughs in this area, even after retirement.
关于作者：周树华，博士，美国阿拉巴马大学传播信息学院教授。主要研究领域有人类对媒体信息的认知，包括信息的表现形式、使用和操纵;认知过程基