Our Methods Reveal Social Structures and Media Effects
by Daniel Kreiss
In their influential article, W. Lance Bennett and Shanto Iyengar trace the history of scholarship on political communication and suggest that we may be moving into a new era of ‘limited effects’ given changes in social structure. Bennett and Iyengar sketch a “premass communication media system” that gave rise to the minimal effects findings in the literature. Social structure in this premass communication system consisted of “relatively dense memberships in a group based society networked through political parties, churches, unions, and service organizations” (707). Accordingly, the authors argue that research findings during this premass era – which the authors never quite define but we can infer lasts up to the early 1960s – related to the two-step flow model and social influence theories more broadly. As these scholars argue about the precursors to the contemporary field as it formed in the 1970s:
“These and other early thinkers [such as Lippmann] all helped position the field of political communication to address the rise of mass society and to grapple with the related understanding of mass communication processes and effects. In this context, the minimal effects and two-step flow models can be explained in retrospect as the result of the studies conducted before the conditions defining mass media and mass society were fully in place. It does not seem particularly surprising that research dating from the 1940s – a time of high social cohesion, before television swept the land, or advertising and polling had become sophisticated – would have produced mixed results about direct attitude change through media messages. Even so, evidence for relatively strong direct effects of political messages in those studies might have warranted more probing analysis.” (715)
Bennett and Iyengar then argue that we see the rise of a “mass society” spanning the 1960s through the 1990s. Research during this era finds “strong media effects” and “direct, mass-mediated, ‘impersonal’ influence processes” (716). Indeed, for these scholars, social structure during this era is characterized by “declining group memberships and the rise of broadcast technologies that made vast audiences accessible via relatively few channels” (716). This sets the stage for their argument that we are now in a new era:
“In addition to the proliferation of channels and fragmentation of the audience, it also makes sense to address in our new political communication models the decline of social conformist identity processes that formerly defined individuals as message receptors in the group membership society that some observers lament losing (Putnam, 2000), along with the decline of the mass audience of “impersonal” social cue takers that defined the mass media social structure (Mutz, 1998; Zaller, 1992).” (716)
Even with the qualifier that early works needed “more probing analysis,” the underlying assumption of the Bennett and Iyengar article is that scholars are actually measuring empirical realities in the world; that research findings were not just an artifact of particular analytical and methodological toolkits marked by different paradigms of communication research. This comes despite Iyengar’s own writings from the early 1980s which argued for more experimental research given the likelihood that cross-sectional surveys were only picking up ‘limited effects’. Even more, other models of both ‘social structure’ and ‘effects’ were always present in the literature. W. Russ Neuman and Lauren Guggenheim, using a citation analysis, show that there were co-existing theories and models of effects (and indeed social structure), and that these are better tracked to different paradigms that have their own analytical and methodological orientations. Scholars who see social contexts as being important, for instance, bring to bear an analytical lens and a set of methods that have continued to find strong interpersonal and group-based communication effects well into the era of ‘mass communication.’ One needs to look no further than studies of diffusion or social networks, or Tamotsu Shibutani’s wonderful, and under-read, study of rumor Improvised News published in 1966.
All of which is to say that it is hard to see the products of theory and method providing, somehow, an unfiltered look at the ontological state of the world. The “methodological pluralism” that has come to dominate the field since the early 1980s – experiments and nationally representative surveys – are simply ill-suited to reveal whether and how social contexts matter in political communication. As such, social contexts have the tendency to disappear from representations of social reality (a point on the performativity of method which John Law forcibly made in “Seeing Like a Survey”; for instance, while social media analyses in recent years have made social ties particularly visible again in political communication research, it is not as if we ever stopped being ‘networked’ or relying on social contacts for our understandings of political life – even if there may be qualitative differences in our social and cultural practices today.) Indeed, even the idea of a ‘mass society’ that posited atomized individuals particularly susceptible to media influence was not uncontested in its time. Raymond Williams famously stated that “there are in fact no masses; there are only ways of seeing people as masses.” But perhaps the most forceful statement is by Daniel Bell in 1960 (who, after all, noted the coming of post-industrial society in 1974):
“What strikes one first about these varied uses of the concept of mass society is how little they reflect or relate to the complex, richly striated social relations of the real world….Behind the theory of social disorganization lies a romantic notion of the past that sees society as having once been made up of small “organic,” close-knit communities.
A great weakness in the theory (of mass society) is its lack of history-mindedness. The transition to a mass society, if it be such, was not effected suddenly, explosively, within a single lifetime, but took generations to mature. In its sociological determinism, the hypothesis overlooks the human capacity for adaptiveness and creativeness, for ingenuity in shaping new social forms. Such new forms may be trade unions whose leaders rise form the ranks…or the persistence under new conditions of ethnic groups and solidarities.”
In other words, there is no reason to believe that the break between ‘premass’ and ‘mass’ was radical. Even more, as Bell argued, that it could be achieved so quickly. Note, for instance, the lovely Middletown study that made a set of empirical claims for the dissolution of group life and local communities in 1929. Indeed, my seminar, ‘A History of the Study of Mass Communication,’ just wrapped up reading what we began collectively referring to as the ‘Communication Bible’ – John Durham Peters and Peter Simonson’s magisterial Mass Communication and American Social Thought. Common tropes that course through work all the way through the 1950s is the dissolution of communities, individualization, and strong media effects – and, the idea of ‘effects’ is far more broadly (and refreshingly) defined than in the contemporary literature (consider Middletown’s claim that once music entered the “machine stage” it “almost ceased to be a matter of spontaneous, active participation and has become largely a passive matter of listening to others”.) For example, in The Public and its Problems, published in 1927, John Dewey argued that there was a proliferation of associations, the dis-embedding of social life, multiplication of publics, and erosion of local, community life:
“One phase of the workings of a technological age, with its unprecedented command of natural energies, while it is implied in what has been said, needs explicit attention. The older publics, in being local communities, largely homogeneous with one another, were also, as the phrase goes, static. They changed, of course, but barring war, catastrophe and great migrations, the modifications were gradual. They proceeded slowly and were largely unperceived by those undergoing them. The newer forces have created mobile and fluctuating associational forms. The common complaints of the disintegration of family life may be placed in evidence. The movement from rural to urban assemblies is also the result and proof of this mobility. Nothing stays long put, not even the associations by which business and industry are carried on. The mania for motion and speed is a symptom of the restless instability of social life, and it operates to intensify the causes from which it springs.”
This seems precisely the social structure that would be conducive to strong media effects. Even more, in 1948 Louis Wirth declared that “In modern urban industrial society, our membership in each of the multiple organizations to which we belong represents our interests only in some limited aspect of our total personal life. There is no group which even remotely professes to speak for us in our total capacity as men or in all of the roles we play.” This was, of course, during the same time period that Paul Lazarsfeld and his colleagues were arguing for the importance and role of social groups and attachments in mediating the effectiveness of mass media; which of course was also a claim that Todd Gitlin famously critiqued on both theoretical and empirical grounds.
All of which suggests that social structures are lumpy (or differentially distributed), historically evolving, and revealed differently according to the analytical and methodological approaches through which we approach the social world. These are not original thoughts. A little more than ten years after Wirth, in 1961, Thelma McCormack argued that social life is differentially structured and pointed to the fact that mass media has varying effects depending on the social context – with traditional communities looking very different from urban environments. Indeed, McCormack’s point is that we cannot look for universal media effects, or generalize from studies in one locale, given that they are conditional on social structures, not just the psychology of individuals. The field would look very different, for instance, if we took seriously Michael Burawoy’s argument that: “Insofar as meaning, attitudes, and even knowledge do not reside with individuals but are constituted in social relations, then we should be sampling from a population of social situations and not a population of individuals.”
This is not to say that social structures do not change, of course. I agree with Bennett and Iyengar that we may be seeing significant shifts in social and media structures over the last few decades as well as different practices for creating social and cultural attachments – but our analytical emphasis and methodological tools also reveal different aspects of the world. Consider the fact that Kurt Lang and Gladys Engel Lang found quite strong media ‘effects’ throughout the 1950s through their groundbreaking empirical research. Through an innovative mixed methods design, their MacArthur Day study of 1952 catalogued how the mass media shaped perceptions of the event for mediated spectators; how anticipatory coverage shaped both the expectations of live spectators and brought them to the event in the first place (a particularly nice recursive finding); and, finally, that the presence of television cameras actually helped create the event itself: “the cheering, waving, and shouting was often largely a response to the aiming of the camera.”
The argument that political communication research tracks changes in social structure means that our theories and method rather unproblematically reveal the empirical world. This idea essentially ignores a number of anomalous works that do not fit the received narrative of limited to strong effects – a point which Neuman and Guggenheim note so forcefully. And, oftentimes these anomalous works provide examples of how different methods reveal not only different social structures but also different kinds of effects. The broad takeaway here is that methods shape how we understand the world and what types of things we think exist, and that there are many kinds of ‘effects’ that have historically been considered in the field of political communication.