Diffusion Theories

Ronald W Perry. Encyclopedia of Sociology. 2nd edition, Volume 1. Macmillan Reference USA, 2001.

The concept of diffusion inherently focuses upon process. Diffusion refers to the dissemination of any physical element, idea, value, social practice, or attitude through and between populations. Diffusion is among the rare concepts used across the physical, natural, and social sciences, as well as in the arts. Diffusion is most closely associated with the social sciences, particularly rural sociology, anthropology, and communication. Diffusion thinking offers a logic through which to describe and perhaps explain myriad types of change that involve equally diverse foci, ranging from the adoption of internet technology (Adams 1997), to the spread of belief systems (Dean 1997).

Work connected to the concept of diffusion is arguably structured as theory. Certainly there is no single, unified, deductively structured collection of propositions, widely regarded by social scientists as identifying the principal mechanisms of diffusion that could be employed across all substantive areas. There are however, many distinct collections of propositions, a few well-tested over decades, that describe different diffusion phenomena in different content areas. Indeed, in the area of innovation diffusion, Everett Rogers (1983) produced a formal theory that is broadly recognized, often tested, and that has been adapted to other content areas including disaster research and technology transfer. Part of the problem in formalizing diffusion theories is that the concept does not inherently specify content (rather a framework or process to structure thinking). However, it cannot be examined empirically without tying it to some substance. That is, studies of diffusion focus on something (a technology, idea, practice, attitude, etc.) that is being diffused. Consequently, the research on diffusion that would drive theory building has remained scattered in the literatures of different sciences, and as such it is not readily pulled together. These conditions do not facilitate ready assembly of information that would encourage creation of a general theory of diffusion. Thus, one must look to the growth of formal theory in sub-areas, although Rogers (in press) has begun to cross content spheres.

There are at least three traditions or theory families that can be historically discerned in the study of diffusion. Rogers (1983, p. 39) has pointed out that for many years these traditions remained largely distinct, with little overlap and cross-fertilization. Since the late 1970s the level of research and theoretical isolation has decreased, leading to an enhanced awareness among the perspectives and some integration of empirical findings into more general theoretical statements. The three theory families are: (1) cultural diffusion; (2) diffusion of innovations; and (3) collective behavior.

Cultural Diffusion

The earliest social scientific use of the term diffusion is found in Edward Tylor’s (1865) treatment of culture change. Anthropologists have long attempted to explain similarities and differences among cultures, especially those that were geographically adjacent. Tylor’s work on culture change first proposed the notion of diffusion as a means of explaining the appearance of similar culture elements in different groups and of understanding the progressive alteration of elements within the same group. As the twentieth century began, diffusion arose as an alternative to evolution as a basis for understanding cultural differences and change. Evolutionists argued that cultural similarities probably arose through independent invention. Those who embraced diffusion presented it as a more parsimonious explanation, emphasizing that traits and institutions could pass between groups by means of contact and interaction.

Historical Development. The English anthropologists W. J. Perry and Elliot Smith devised the most extreme position on cultural diffusion. These scholars held that human culture originated in Egypt and progressively diffused from that center over the remainder of the earth. In Germany, Fritz Graebner (1911) argued that critical aspects of cultures—toolmaking, for example—originated in a small number of geographically isolated societies. This hypothesis formed the basis for culture circles (“kulturkreise”), collections of societies sharing similar cultures. Unlike British diffusionists who emphasized tracking the movement of single culture elements, Graebner and others in his tradition focused on the dissemination of collections of elements or cultural complexes.

American anthropologists are credited with developing a social scientifically workable concept of diffusion. Franz Boas (1896) conceived of diffusion as a viable mechanism for culture exchanges among geographically adjacent areas. His view figured prominently in the intellectual move away from the deterministic view of diffusion proposed by early British anthropologists. Alfred Kroeber (1923, p. 126) and Robert Lowie (1937, p. 58)—students of Boas—subsequently developed a position called moderate diffusionism, which is currently widely accepted in anthropology. This position allowed for the coexistence of a variety of mechanisms of change and transfer—independent invention, acculturation, etc.—in addition to diffusion in accounting for culture change and differentiation. Clark Wissler (1929), a Kroeber contemporary, established an empirical basis for culture diffusion by identifying ten culture areas (regions with similar cultural inventories) in North and South America and the Caribbean.

Theory in Cultural Diffusion. In terms of theory development, cultural diffusion is the actual movement of a given social institution or physical implement, while stimulus diffusion is the exchange or movement of the principle upon which an institution or implement is based. In the cultural diffusion literature, scholars have enumerated assumptions, stated principles, and reviewed empirical work with the objective of identifying propositions tested repeatedly and not found to be false. Indeed, beginning with the work of early twentieth century anthropologists, one can identify at least five broadly accepted and empirically supported claims that form the core of what is called cultural diffusion theory. First, borrowed elements usually undergo some type of alteration or adaptation in the new host culture. Second, the act of borrowing depends on the extent to which the element can be integrated into the belief system of the new culture. Third, elements that are incompatible with the new culture’s prevailing normative structure or religious belief system are likely to be rejected. Fourth, acceptance of an element depends upon its utility for the borrower. Finally, cultures with a history of past borrowing are more likely to borrow in the future. These claims constitute the “core propositions” of culture diffusion theory; over the years, each has been qualified and elaborated upon, and corollaries have been created (Stahl 1994).

Currently, diffusion is seen as a mechanism for culture change that typically accounts for a large proportion of any particular culture inventory. The deterministic, linear view of diffusion has been discredited by the empirical record. The concept of culture diffusion as a means of understanding cultural inventories is entrenched in the field of cultural anthropology. As the twenty-first century dawns, a principal controversy among cultural anthropologists centers on the definition of culture, rather than upon the acceptability of culture diffusion. Diffusion theory remains prominent in the archaeology literature, particularly as a means of tracing culture inventories for groups over time (Posnansky and DeCorse 1986).

Sociologists were initially involved in the use of cultural diffusion theory as a means of looking at cultural change (largely in terms of nonmaterial culture) in the United States. Initially theory drawn from anthropology was used (Chapin 1928), but over time the sociological focus became identifying social psychological motivations and mechanisms supporting the diffusion process (Park and Burgess 1921, p. 20). Recently, sociological work directly on culture is bifurcated, with one group of scientists still emphasizing the social psychological issues in culture meaning (Wuthnow and Witten 1988), and another more concerned with structural (mathematical or statistical) models of culture processes themselves (Griswold 1987). Neither group has especially focused on diffusion theory as a mechanism to track or identify the content-outcomes of culture change.

Diffusion of Innovations

The diffusion of innovations has historically focused on the spread of an idea, procedure, or implement within a single social group or between multiple groups. For the most part, scholars of this tradition define diffusion as the process through which some innovation is communicated within a social system. Also important is the notion of a time dimension reflecting the rate of diffusion, and the importance of the individual adopter (or non-adopter) reflecting the role of social influence.

The study of innovation diffusion began rather narrowly, grew to dominate the field of rural sociology for a time, contracted in popularity for many years, and then spawned wide interest across several disciplines. Innovation diffusion study contains several groups: those who focus on content or the specific innovation being diffused; those who emphasize theoretical elaborations of generic principles of innovation diffusion; and those concerned with creating structural models to track diffusion. Particularly in the past decade, the literature has seen much cross-fertilization, although mathematical modelers tend to appear less often in the work of other diffusion scholars. Although the roots of innovation diffusion theory are seen to be largely in rural sociology, more recently the field has become distinctly interdisciplinary with major advancements made especially in the discipline of communication.

Historical Development. The definitive history of the diffusion of innovations as a paradigm was published by Thomas Valente and Everett Rogers (1995). The roots of innovation diffusion are usually traced to Gabriel Tarde (1890) who didn’t use the term diffusion, but was the first to address the notions of adopters and the role of social influence in adoption, as well as to identify the S-shaped curve associated with the rate of an innovation’s adoption. The formative empirical work on innovation diffusion can be traced to Bryce Ryan’s Iowa State University-based study of hybrid corn seeds (published with Gross in 1943), and Raymond Bowers’s (1937) study of the acceptance and use of ham radio sets. For more than two decades following this pioneering work, the study of innovation diffusion and particularly theory development took place within the context of rural sociology. This circumstance was a function of a variety of forces, principal among which were the location of rural sociologists in land grant institutions charged with the dissemination of agricultural innovations to farmers (Hightower 1972) and the communication and stimulation accorded by the North Central Rural Sociology Committee’s (a regional professional society) formation of a special subcommittee to deal with the issue of diffusion of agricultural innovations (Valente and Rogers 1995, p. 254).

Most scholars agree that contemporary views of innovation diffusion grew from hybrid corn seeds; specifically the research on adoption done by Bryce Ryan and Neal Gross (1943). These studies ultimately defined most of the issues that occupied diffusion researchers and builders of innovation diffusion theory for decades to come: the role of social influence, the timing of adoptions, the adoption process itself, and interactions among adopter characteristics and perceived characteristics of the innovation. From the middle 1940s through the 1950s, rural sociologists vigorously developed a body of empirical information on the diffusion of innovations. Most of these studies remained tied to agriculture and farming, and focused on the diffusion of new crop management systems, hybridizations, weed sprays, insect management strategies, chemical fertilizers, and machinery. A common criticism of the studies of this era is that many of the studies seem to be almost replications of the Ryan and Gross work, the main difference among them being the specific innovation studied. While it is true that these studies tend to share a common methodology and linear conception of diffusion, it is also true that they provide a strong foundation of empirical case studies. Indeed, the replications that these studies represent substantially facilitated the later sophisticated theoretical work initiated in the early 1960s (Rogers 1962), and continued in the 1980s (Rogers 1983, 1988).

The 1960s marked the beginning of the decline of the central role of rural sociologists in innovation diffusion research. In large part this was due to changes in the field of rural sociology, but it also reflected the increasing involvement of researchers from other disciplines, changing the sheer proportion of rural sociologists working on innovation diffusion. After more than two decades of extensive research on the diffusion of agricultural innovations, rural sociologists—like other social scientists of the time—began to devote more time to the study of social problems and the consequences of technology. Indeed, Crane (1972) argued that around 1960 rural sociologists began to believe that the critical questions about innovation diffusion had already been answered. Although the late 1960s saw rural sociologists launch a series of diffusion studies on agricultural change in the international arena (particularly Latin America, Asia and Africa), by 1965 research on diffusion of innovations was no longer dominated by members of that field. Of course, innovation diffusion research by rural sociologists has continued, including studies of the impacts of technological innovation diffusion, and diffusion of conservation practices and other ecologically-based innovations (Fliegel 1993).

The infusion of researchers from many disciplines studying a variety of specific innovations initiated the process of expanding the empirical testing of innovation diffusion tenets. This began with studies in education addressing the diffusion of kindergartens and driver education classes in the 1950s, as well as Richard Carlson’s (1965) study of the diffusion of modern math. Another major contribution came from the area of public health. Elihu Katz, Herbert Menzel, and James Coleman launched extensive studies of the diffusion of a new drug (the antibiotic tetracycline); first in a pilot study (Menzel and Katz 1955) and then in studies of four Illinois cities (Coleman, Menzel, and Katz 1957; Coleman, Katz, and Menzel 1966). This research greatly expanded knowledge of interpersonal diffusion networks, and in particular its influence in adoption. Interestingly, as Elihu Katz, M. L. Levine, and Harry Hamilton (1963) indicated, the drug studies truly represented an independent replication of the principles of innovation diffusion developed by rural sociologists because the public health researchers were unaware of the agricultural diffusion research. Other studies in the public health arena focused on dissemination of new vaccines, family planning, and new medical technology.

Beginning in the late 1960s, there was a substantial increase in the amount of diffusion research in three disciplines: business marketing, communication, and transportation-technology transfer. Marketing research principally addressed the characteristics of adopters of new products and the role of opinion leaders in the adoption process (Howes 1996). This literature is based almost exclusively on commercial products, ranging from coffee brands and soap to touch-tone telephones, the personal computer, and internet services. The studies tend to be largely atheoretical, methodologically similar, and aimed simply at using knowledge of diffusion either to improve marketing and sales of the product or to describe product dissemination.

In sharp contrast, work done on innovation diffusion by scholars trained in communication has been considerably more theoretically oriented. Throughout the 1960s, universities in America began to establish separate departments of communication (Rogers 1994). Since diffusion of innovations was widely seen as one type of communication process, scholars in these new departments adopted this type of research as one staple of their work. Beginning with studies of the diffusion of news events (Deutschmann and Danielson 1960), this research tradition has branched out to study the dissemination of a wide variety of specific innovations (McQuail 1983, p. 194). Scholars working in this tradition have been principally responsible for the progressive refinements of formalized theory of innovation diffusion. Everett Rogers has consistently remained the leader in theory development in communication, revising and extending his 1962 book Diffusion of Innovations with help from co-author Floyd Shoemaker to produce Communication of Innovations in 1971. Subsequently, Rogers (1983, in press) restored the original title Diffusion of Innovations, broadened the theoretical base and incorporated diffusion studies and thinking from other disciplines. Generally, Rogers and other communication scholars have studied the diffusion of many target material elements, phenomena, and other intangibles, but they have continued to produce theoretical statements dealing with communication channels, diffusion networks, interpersonal influence and the innovation-decision process. Finally, the technology dissemination and transfer issues have involved work by geographers, engineers, and others beginning in the 1970s. The primary focus of such studies has been the spread or dissemination of technology (Sahal 1981) and the development of network models of innovation diffusion (Valente 1995).

Theory in Innovation Diffusion. The theoretical work of Everett Rogers initially resulted in the collection of knowledge gained from the rural sociology tradition, then facilitated the transition to communication perspectives, and now has served as the mainstay of what is developing as a more cross-disciplinary focus on innovation diffusion. His contribution is twofold. First, he created inventories of findings from many disciplines and from many types of innovation. These inventories provided impetus for the development of a definition of innovation diffusion that was not bound by discipline. Second, Rogers assembled and refined theoretical structures aimed at explaining the principal features of innovation diffusion. The theoretical work has cemented a core of knowledge and principles that are widely identified (and used empirically) as the bases of the diffusion of innovations. Rogers’s (1983) theory includes eighty-one generalizations (propositions) that have undergone empirical testing.

The theory of innovation diffusion may be understood as capturing the innovation-decision process, innovation characteristics, adopter characteristics, and opinion leadership. The innovation-decision process represents the framework on which diffusion research is built. It delineates the process through which a decision maker (representing any unit of analysis) chooses to adopt, reinvent (modify), or reject an innovation. This process consists of five stages. Knowledge is the initial stage when the decision maker detects the existence of the innovation and learns of its function. In the persuasion stage, the decision maker forms a positive or negative attitude toward the innovation. The third stage, decision, deals with the decision-maker’s choice to accept or reject the innovation. Implementation, the fourth stage, follows a decision to accept and involves putting the innovation into some use (in either its accepted form or some modified form). During the final stage of confirmation, decision makers assess an adopted innovation, gather information from significant others, and choose to continue to use the innovation as is, modify it (reinvention), or reject it. While some have criticized the stage model as too linear, Rogers (1983) has convincingly argued that existing formulations afford a degree of interpretative and predictive flexibility that averts historical problems with stage models in social science.

Different innovations have different probabilities of adoption and hence, different adoption rates. That is, they travel through the innovation-decision process at varying speeds. The literature demonstrates that five characteristics of innovations influence the adoption decision. Compatibility refers to the congruence between an innovation and the prevailing norms, values, and perceived needs of the potential adopter. Higher levels of compatibility are associated with greater likelihood of adoption. Innovation complexity, on the other hand, is negatively associated with adoption. The extent to which use of an innovation is visible to the social group—called observability—is positively related to adoption. Relative advantage refers to the extent to which an innovation is perceived to be “better” than the idea, practice, or element that it replaces. Higher relative advantage increases the probability of adoption. Finally, trialability—the extent to which an innovation may be experimented with—also increases the probability of adoption.

The third component of diffusion of innovation theory addresses adopter characteristics. Adopter categories are classifications of individuals by how readily they adopt an innovation. Rogers (1983, p. 260) identifies nine socioeconomic variables, twelve personality variables, and ten personal communication characteristics that have been demonstrated to bear upon adoption choices. In general, the literature holds that early adopters are more likely to be characterized by high socioeconomic status, high tolerance of uncertainty and change, low levels of fatalism and dogmatism, high integration into the social system, high exposure to mass media and interpersonal communication channels, and frequent engagement in information seeking.

Identifying the characteristics of people who adopt innovations raises the question of interpersonal influence. Three issues are addressed in the development of propositions about the role of interpersonal influence in the innovation decision process: information flow, opinion leadership, and diffusion networks. Over time, information flow has been seen as a “hypodermic needle” model, a two-step flow (to opinion leaders, then other adopters), and a multi-step flow. Currently, information flows are seen as multi-step in nature and are described in terms of homophily and heterophily—the degree to which pairs of interacting potential adopters are similar or dissimilar. Opinion leadership denotes the degree to which one member of a social system can influence the attitude and behavior of others. This concept is presently discussed relative to spheres of influence, wherein a given person may be a leader or follower depending upon the part of the diffusion network being referenced. The diffusion or communication network is the structural stage upon which social influence takes place. Considerable attention has been devoted to developing analysis strategies and tactics for such networks (Wigand 1988).

Collective Behavior

While diffusion is not a commonly used term in collective behavior, processes of diffusion are important in connection with understanding crowds, fashion, and some aspects of disaster behavior. In all cases, analytic concern centers on the dissemination of emotions, social practices, or physical elements through a collectivity. The study of human behavior in disasters is recent and multidisciplinary. In this field there has been a concern with diffusion in the classic sense of tracking ideas and practices through networks. The principle foci of research have been the adoption of protective measures and the dissemination of warning messages (Lindell and Perry 1992), with the aim of research being both the development of general theories of protective behaviors and more effective protection of endangered populations.

All three diffusion theory traditions converge in the study of crowd behavior. In proposing imitation as an explanatory mechanism for crowd actions, Gabriel Tarde (1890, p. 45) drew upon Edward Tylor’s concept of cultural diffusion. Subsequently, Gustave LeBon (1895) and Gabriel Tarde (1901) approached crowd behavior in terms of social contagion: rapid dissemination of emotions among interacting people. Although Floyd Allport (1924) and Herbert Blumer (1939) extended and formalized the concept of contagion, it has been largely displaced as a theory of crowd behavior by convergence theory (Turner and Killian 1987, p.19).

Changes in dress have been conceptualized as diffusion processes. Alfred Kroeber (1919) studied fashion cycles which he believed diffused systematically through civilizations. Katz and Lazarsfeld (1955, p. 241) moved away from the initial concern with movement of fashion through networks to focus more on social influence. Herbert Blumer (1969) firmly established social psychological motivations as the basis for fashion behavior. Current theoretical work on fashion continues to emphasize social psychological approaches wherein fashion diffusion issues are peripheral (Davis 1985; Nagasawa, Hutton, and Kaiser 1991).