28 IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, VOL. EM-29, NO. 1, FEBRUARY 1982  Innovation Characteristics and Innovation Adoption- Implementation: A Meta-Analysis of Findings  LOUIS G. TORNATZKY AND KATHERINE J. KLEIN Abstract—A review and meta-analysis was performed of seventy- five articles concerned with innovation characteristics and their  relationship to innovation adoption and implementation. One part of  the analysis consisted of constructing a methodological profile of the  existing studies, and constrasting this with a hypothetical optimal  approach. A second part of the study employed meta-analytic sta- tistical techniques to assess the generality and consistency of existing  empirical findings. Three innovation characteristics (compatibility,  relative advantage, and complexity) had the most consistent sig- nificant relationships to innovation adoption. Suggestions for future  research in the area were made.  INTRODUCTION I NNOVATION characteristics research describes the relation- ship between the attributes or characteristics of an innova- tion and the adoption or implementaion of that innovation. This topic represents one of the classic issues in the innovation literature albeit one that has been little studied in the last decade. As an area of considerable past activity, it has been reviewed (Rogers and Shoemaker, 1971; Rothman, 1974; Zaltman and Lin, 1970) and critiqued (Downs and Mohr, 1976) previously. While both Rogers and Shoemaker's and Rothman's reviews are quite comprehensive, their respective analyses of the literature largely exclude any detailed examin- ation of the reviewed studies' methodological or conceptual rigor. But, nonetheless, it is on the basis of such general (not detailed methodological) reviews that others, especially Downs and Mohr, have criticized innovation characteristics research as a body of inquiry. This paper presents the results of a comprehensive litera- ture review and preliminary meta-analysis of studies of inno- vation characteristics. In order to supplement earlier reviews, we provide a methodological and conceptual profile of innova- tion characteristics literature. The study is also designed to ex- amine the extent to which Downs and Mohr's criticisms are appropriate. Finally, we provide our own recommendations for the future study of innovation characteristics. Before turning to the methods, results, and discussion of this meta-analysis, further discussion of Downs and Morns' criticism of the innovation characteristics literature and also our own observations for the ideal innovation characteristic study are in order. Manuscript received March 2, 1981; revised July 16, 1981. L. G. Tornatzky is with the Division of Industrial Science and Tech- nological Innovation, National Science Foundation, Washington, DC. K. J. Klein is with the Department of Psychology, University of Texas, Austin, TX. The Downs and Mohr Critique In their review of the innovation literature, Downs and Mohr (1976) devote considerable attention to what they see as fatal flaws in the studies of innovation characteristics. According to Downs and Mohr, this body of research focuses on both pri- mary and secondary attributes of innovations but fails to pay sufficient attention to the distinction between the two. Pri- mary attributes are seen as inherent to the innovation or tech- nology and invariant across settings and organizations; second- ary attributes are defined as perceptually based (or subjective) characteristics. An example of a so-called primary attribute is size or cost. Complexity or relative advantage are secondary attributes. Perception of secondary attributes is assumed to be influenced by characteristics of both the particular setting and the actors involved in implementing a particular innovation. If anything, Downs and Mohr (1976) probably underpay the importance of "subjective" factors. Downs and Mohr ig- nore the perceptual literature in social psychology and related fields which has for many years noted that even what is as- sumed to be invariant physical reality (e.g., a primary attri- bute) is always subject to social influences. For example, group conformity studies demonstrated that respondents make startling alterations in perceptions of physical dimensions (Asch. 1956) when confronted with a synthetic majority opin- ion to the contrary. Furthermore, while so-called primary attributes of innova- tions can be measured "objectively," the meaning of the ob- jective measure of the characteristic is subjective, that is, in the mind of the perceiver. Thus while an innovation may cost a fixed amount (and cost is a so-called primary attribute), the cost of the innovation is evaluated by the potential adopter relative to his or her financial resources. The innovation's cost may seem inexpensive to one, exorbitant to another. In this sense, there can be no primary attribute of an innovation. Perceptions are always evaluated in reference to some inter- nalized system of values or cognitive framework; the result is a subjective rating of the significance of the "fact" (e.g., size, cost, etc.) In any case, Downs and Mohr do rightly point out that "most if not all characteristics upon which one might con- sider basing a typology turn out to be secondary attributes of innovations" (p. 703). More important for this discussion than the distinction be- tween primary and secondary attributes per se is the inference that these authors make as a result. Downs and Mohr's conclu- sion is that "because an innovation's classification in a second- ary-attributed topology can vary from organization to organi- 0018-9391/82/0200-0028$00.75 © 1982 IEEE