|
Devising appropriate measures of adopter innovativeness is one of the more important tasks in the innovation diffusion field. Many measures have been proposed, including traditional
measures, such as time of adoption, and newer measures, such as stage of assimilation. This study provides both a conceptual and an empirical analysis of six prominent measures of the adopter innovativeness
concept: 1) Time of Adoption, 2) Dichotomous Adoption, 3) Aggregated Adoption, 4) Internal Diffusion, 5) Infusion, and 6) Assimilation Stage. The conceptual analysis describes the origins of each measure,
identifies potential advantages and limitations, and provides guidance on when and how the measure is likely best applied. The empirical analysis uses data on the adoption of three software process
innovations (object-oriented programming languages, relational database management systems, and CASE tools) to operationalize the six measures and to explore several issues raised in the conceptual review. The conceptual analysis showed that each measure had a unique set of potential strengths and weaknesses, and that no measure dominated any of the
others. The empirical analysis provided several additional insights. It was shown that: 1) even thin measures such as Time of Adoption and Dichotomous Adoption can have reasonable strong criterion validity;
2) contrary to the concern first raised by Downs and Mohr, there are positive relationships between the propensities to acquire many innovations, to acquire innovations early, and to implement innovations in
depth; 3) the use of Aggregated Adoption has a strong positive effect on variance explained, and leads to a pattern of significant relationships identical to that predicted by theory, and 4) the Extent of
Implementation measures (Internal Diffusion and Infusion) have very weak criterion validity, quite possibly due to restriction in the ranges of study variables.Overall, it appears that, with
the possible exceptions of Internal Diffusion and Infusion, the alternative innovativeness measures can be seen as valid indicators of a general notion of adopter innovativeness. As a result, researchers
have the freedom to let their substantive interest in different classes of predictors drive the research design process, while still maintaining a strong link to the cumulative tradition of diffusion
research. |
