The present study proposes a model of agenda diffusion as an explanation for inconsistencies in agenda-setting research. We suggest that interpersonal communication plays a pivotal role in a two-step process in which the media agenda diffuses first from the mass media to media users, and then from these via interpersonal communication to non-users. We present findings from a field study to provide empirical evidence for the reality of agenda diffusion and discuss theoretical, analytical, and methodological implications for further agenda-setting research.
Since the seminal study by McCombs and Shaw,1 researchers have accumulated strong evidence for agenda setting in different contexts, and the agenda-setting effect of mass media has been well established in the theoretical repertoire of communication and political science. A recent content analysis of the media effects literature showed that agenda setting is one of the three theories that are most often referred to in publications about mass media effects.2 However, there is a puzzling inconsistency between studies analyzing agenda setting on the macrosocietal level and studies on individual data level. Whereas studies at the macro-level often found large agenda-setting effects, micro-level studies often found weak effects or no effect at all. The basic assumption of agenda setting, however, refers to cognitive processes within individuals: as a consequence of frequent and prominent coverage of certain issues by the mass media, members of the public perceive these issues as more important than others.
This article seeks to propose an explanation for this discrepancy by introducing a model of agenda diffusion as a process of mass media effects and interpersonal communication. It integrates ideas of the twostep flow of mass communication,3 diffusion of information and innovation,4 and campaigns and conversation5 into the agenda-setting hypothesis.
We first discuss the aforementioned inconsistency in media effect studies and present an explanation of how interpersonal communication could be a key factor to clear up the ambiguity of previous findings. Second, we review how interpersonal communication has been examined in agenda-setting research to date and point out indications for our assumptions about the role of interpersonal communication in agenda setting. Finally, we introduce our agenda-diffusion model and present findings from a field study to test our hypotheses.
Inconsistencies in Media Effects Research
In media effect studies, we often find a paradox: studies examining media effects on the aggregate level often found large media effects, whereas studies examining effects on individuals mostly found only weak effects or no media effect at all.6 We also find this paradox in agenda-setting research when comparing aggregated and individual level studies.7 In his review, Weaver8 found systematically greater effects in agenda-setting studies on the aggregated level than on the individual level.9 The greatest differences were found between high rank correlations between aggregated media agenda and aggregated public agenda10 on the one hand and weak correlations between the individual agenda and the agenda of the individual media input on the other hand. This result is surprising, as the theoretical assumption underlying agenda setting refers to the individual level: the individual is told by the press (or TV) what to think about.11 Hovland12 argues that individual media effects in non-experimental settings are accompanied by "supplementary effects produced by discussion with friends." In consequence, conversation may limit the individual media effects by giving competitive information,13 triggering resistant social norms,14 or activating inoculation.15
We think that the relationship between agenda-setting effects of the mass media and interpersonal communication might be part of the explanation for this discrepancy. Recall the idea of the paradox of media effects on non-users. …