Diffusion Theory in the New Media Environment: Toward an Integrated Technology Adoption Model

Abstract

This article explores the venerable diffusion of innovations model and how changing technologies impact its applications and generalizations viewed as “products” of the model. We've examined some of the concepts involved, including characteristics of innovations, stages in the process, and characteristics of adopters. Then we attempted to develop a more complex understanding of the diffusion process by integrating theoretical frameworks from information sciences and uses and gratifications theory in developing a model for adopting technologies themselves.

Notes

¹Rogers (2003) noted of diffusion research that “no other field of behavior science research represents more effort by more scholars in more disciplines in more nations” (p. xv). Although the diffusion framework is synonymous with Rogers's work, the historical underpinnings of diffusion can be linked to the European scholar, Gabriel Tarde (Rogers, 2003). The first empirical study of diffusion involved the hybrid corn seed study in Iowa by Ryan and Gross (Rogers et al., 2009).

²Chaffee and Berger (1987) offered seven criteria by which the quality of a theory can be judged: explanatory power, predictive power, parsimony, internal consistency, heuristic provocativeness, organizing power, and falsifiability.

³Diffusion research thus addresses a number of new communication technologies, including cable television (e.g., Krugman, 1985; LaRose & Atkin, 1988), high-definition television (e.g., Dupagne, 1999), interactive television (e.g., Leung & Wei, 1998), digital television (e.g., Atkin, Neuendorf, Jeffres, & Skalski, 2003), audiotext (Lin & Atkin, 2002), personal computers (e.g., Dutton et al., 1987; Lin, 1998), the Internet (e.g., Atkin, Jeffres, & Neuendorf, 1998; Rogers, 2002), mobile phones (e.g., Leung & Wei, 1999, 2000), webcasting (e.g., Cha & Chan-Olmsted, 2012; Lin, 2004), online radio (e.g., Lin, 2009), and social media (Hunt, Atkin, & Krishnan, 2012, 2014a, 2014b; Webster, 2010).

⁴Not all diffusion studies have operationalized innovativeness as a composite of the adopter categories. Diffusion research has thus been criticized for its lack of a universal measurement for the innovativeness construct (Goldsmith & Hofacker, 1991). Goldsmith and Foxall (2003) explained that variations in the way innovativeness is conceptualized often depend on a given researcher's goals. Hurt et al. (1977) commented that many conceptualizations of innovativeness treat the construct as innovation specific; if any innovation is not appropriate to the target group, it will lead to nonadoption. The authors created a scale that would treat innovativeness as a personality characteristic that could be administered as a self-report survey. Midgley and Dowling (1978) commented that diffusion research has measured innovativeness in one of two ways. The first approach is to operationalize innovativeness related to the time of adoption, such as the individuals who purchase a product in a designated amount of time; the other involves measuring similar product purchases or the adoption of technology clusters.

In particular, Midgely and Dowling (1978) specified the importance of distinguishing between “innate innovativeness” and “actualized innovativeness.” According to the authors, innate innovativeness is a personality trait or an inferred attribute variable, whereas actualized innovativeness is what can be observed and measured. Lin (1998) operationalized the “need for innovativeness” construct because it allowed for the important distinction between “innate” and “actualized” innovativeness. She explained that actual adoption results from the combination of need for innovativeness with other factors such as financial resources and optimal adoption timing. Later research (e.g., Cha & Chan-Olmsted, 2012; Vishwanath & Barnett, 2011) has incorporated these innovativeness scales or measures as alternatives to Rogers's (2003) conceptualization to study the diffusion of different new technologies.

⁵Indeed, Rogers (2003) is notable for providing perhaps the most pointed criticisms of the diffusion framework with which he had become heavily identified. Among those that have received the most attention in the literature are the pro-innovation bias, the individual blame bias, the recall problem, and the issue of equality. The pro-innovation bias is the notion that an innovation should be adopted by all of the members of a social system (Rogers, 2003). The individual blame bias occurs when individual differences are considered as the basis for nonadoption (Rogers, 2003). Another criticism of diffusion research (Rogers, 2003) involves the recall problem among respondents, one that results from respondent inability to recall when they chose to adopt an innovation. Finally, Rogers (2003) noted that diffusion theory is wider than it is deep. That is, across thousands of studies, diffusion research provides an inconsistent gauge for actual adoption.

⁶The limitations of the unified theory of acceptance and use of technology model have also played a role in its lack of acceptance among diffusion researchers. For instance, Peters (2011) maintained that the theory's generalizability is limited outside of organizational adoption contexts—where it was originally applied—due to the quality of constituent constructs. Park (2010) specified several limitations of the model. The first limitation is that the four determinants and four moderators specified in the model may interact with one another. A second limitation is that the model confuses self-efficacy and perceived behavioral control. Other limitations identified by Park include how the model lacks parsimony and misses the unique theoretical strengths of the specific diffusion models. Based on these limitations, Park (2010) argued that combining the technology acceptance model with uses and gratifications theory creates a more suitable framework for adoption research. We concur that certain components of the model are not relevant to new media—and social media research outside of organizational contexts—which are largely determined by audience use motivations.

⁷Recent technology adoption research has combined motives with perceived technology attributes to explain frequency of using interactive communication technologies (e.g., Hunt et al., 2014a, 2014b; Park, 2010). This work also extends to the adoption and use of social media (e.g., Hunt et al., 2012), YouTube watching (e.g., Hanson & Haridakis, 2008), and the like. A typical set of gratifications for Internet-based technology use includes interpersonal utility (e.g., to belong to a group), pass time, information seeking, convenience (e.g., easier to email than tell people), and entertainment (e.g., Charney & Greenberg, 2002; Papacharissi & Rubin, 2000; Sundar & Limperos, 2013).

⁸LaRose and Atkin (1991) mathematically modeled the process of audience adoption and use of established and emerging movie delivery modalities. They found empirical support for a formula of media selection, one that draws from expectancy value theory (Azjen & Fishbein, 1980) to gauge functional displacement among choices. Use of functionally similar technologies thus predicts adoption of myriad media technologies, wherein use of a similar modality presupposes that the utility and ease of use for a given media innovation (Vishwanath & Goldhaber, 2003). Such conceptions can be informed by similar work in other disciplines. For instance, these functional similarity conceptions approximate those found in computer science governing a similarity measure or similarity function. These similarity conceptions specify a real-valued function that quantifies the similarity between two objects (e.g., Frey & Dueck, 2007). Although no single definition of a similarity measure exists, usually similarity measures are in some sense the inverse negative value for very dissimilar objects.

⁹Social capital is often described as the connections we have to others that help reach shared goals and facilitate civic engagement (e.g., Putnam, 2000). Interactive communication platforms, like Facebook, are often designed around such connections. Social networking sites allow users to reach out to their network for help during acts of relationship maintenance, which allows them to benefit from social capital resources of their network (Ellison, Gray, Lampe, & Fiore, 2014). Online users have been found to be higher in their degree of sociability and social norms than nonusers (Lee & Lee, 2010). Information seeking via social networking sites has been demonstrated to increase social capital and civic engagement (Gil de Zúñiga et al., 2012). In an effort to cross communication subdisciplines barriers (mass/interpersonal/organizational) and due to the nature of new media, scholars have conceptualized communication capital as a more appropriate concept (Jeffres, Jian, & Yoon, 2013).

Additional information

Notes on contributors

David J. Atkin

David J. Atkin (Ph.D., Michigan State University, 1988) is a Professor in the Department of Communication at the University of Connecticut. His interests include digital media policy as well as the adoption and uses of emerging media.

Daniel S. Hunt

Daniel S. Hunt (Ph.D., University of Connecticut, 2012) is an assistant professor in the Department of Communication at Worcester State University. His research interests include new media technology, social media, and visual communication.

Carolyn A. Lin

Carolyn A. Lin (Ph.D., Michigan State University, 1987) is a Professor in the Department of Communication at the University of Connecticut. Her research focuses on the content, uses and effects of digital technologies, health communication/informatics, advertising, social marketing and cross-cultural communication.