|
COPYRIGHT 2001 Institute for Operations Research and the Management Sciences
The Technology Acceptance Model (TAM) has received considerable research attention in the IS field over the past decade, placing an emphasis on the roles played by perceived ease-of-use and perceived usefulness in influencing technology adoption decisions. Meanwhile, alternative sets of antecedents to adoption have received less attention. In this paper, sets of antecedent constructs drawn from both TAM and the Perceived Characteristics of Innovating (PCI) inventory are tested and subsequently compared with one another. The comparison is done in the context of a large-scale market trial of a smart card-based electronic payment system being evaluated by a group of retailers and merchants. The PCI set of antecedents explains substantially more variance than does TAM, while also providing managers with more detailed information regarding the antecedents driving technology innovation adoption.
(TAM; PCI; Adoption; Managers; Perceptions; Attitudes; Intentions; Field Study; High Technology; Smart Cards)
**********
Over the past decade, researchers within the information systems (IS) community have sought to conceptualize, empirically validate, and extend various models of individual-level information technology adoption and usage. These models have generally attempted to use key antecedent attitudinal constructs drawn from established psychological theories to predict new IS technology adoption. For example, the Technology Acceptance Model (TAM) proposed by Davis et al. (1989) to explain IS adoption in a variety of contexts incorporates Fishbein and Ajzen's Theory of Reasoned Action (1975) as its theoretical foundation.
TAM has become one of the most widely applied individual-level technology adoption models in the IS literature. Several alternative models of technology adoption have been proposed in an attempt to overcome the limitations of TAM by incorporating additional constructs suggested by theories other than the Theory of Reasoned Action (TRA). For example, Mathieson (1991) proposed a model of technology adoption premised on Ajzen's (1991) Theory of Planned Behavior that expands TAM to include two additional constructs. Other efforts have sought to develop measures of actual system acceptance as opposed to intended usage (Szajna 1996), and to identify important additional antecedent constructs that also underlie the technology adoption decision, such as computer self-efficacy (Campeau and Higgins 1995) and the role of prior experience (Taylor and Todd 1995a).
Although these conceptual and empirical advances have helped IS researchers to better understand the antecedents to technology adoption, at least two broad concerns remain. First, TAM is often employed because of its parsimony and robustness, allowing the user to explain considerable variance while using only two antecedents (perceived usefulness and perceived ease-of-use). However, although parsimony is an important consideration, individual responses to new technologies are likely to differ depending on the context within which they are encountered. Complete understanding of adoption behavior requires a model that captures the richness of the adoption process across many different contexts.
Second, much of the existing IS adoption literature has focused on the adoption of new behaviors, such as the usage of a personal computer or a particular software package. In these studies, subjects are typically asked to assess innovations that are described across a limited range of possible benefits, and to make adoption decisions that involve minimal acquisition costs. Fewer field-based tests of technology adoption models have been undertaken, although their number is growing (for one exception in a marketing context, see Taylor and Todd 1995b). This latter stream of research is important because individuals and firms making decisions in the field must augment their concerns about price/performance issues with more qualitative assessments of image and visibility. Furthermore, the costs of adoption are typically quite large in these settings. Thus, to assess the generalizability of the proposed technology adoption models, it is important to study their application across both experimental and field-based settings.
In this paper, we assess the value of including a wider variety of adoption antecedents to predict technology adoption. In a head-to-head study, we compare TAM's antecedents to the Perceived Characteristics of Innovating (PCI) antecedents developed by Moore and Benbasat (1991). Using Rogers' Diffusion of Innovations theory (1995), Moore and Benbasat developed a robust, reliable, and valid set of 8 constructs that are key antecedents to technology adoption decisions. Despite their attractive conceptual and measurement properties, however, relatively little empirical use has been made of the PCI constructs (for notable exceptions, see Agarwal and Prasad 1997, 1998; Chin and Gapal 1995, Gagliardi and Campeau 1995, Moore and Benbasat 1994). No previous study has directly compared the performance of these two models.
The TAM and PCI antecedents are examined in the context of an ongoing market trial of a smart card payment system--a point-of-purchase technology that electronically conducts and manages purchase transactions. Specifically, we examine retailer and merchant interest in adopting this new technology for use in their own businesses. Our primary purpose in conducting this study is to determine which model provides a more complete explanation of variance in the context of a real-world adoption decision. We find that when used as antecedents to adoption intention, the PCI belief constructs explain a higher proportion of the variance than TAM. Thus, our results suggest that there is value in sacrificing parsimony to include a richer set of antecedents to predict adoption.
The remainder of the paper is organized as follows. First, TAM and the PCI belief constructs are briefly reviewed. Next, the empirical setting in which these models are tested is described, along with a discussion of the sample, research method employed, and results. This is followed by a discussion of the study's key findings, along with its broader implications for IS technology adoption research. Finally, the paper concludes by acknowledging the key limitations of the study, as well as offering suggestions for additional research.
Literature Review
The Technology Acceptance Model (TAM)
TAM builds primarily from the theory of reasoned action (Fishbein and Ajzen 1975) but also draws from expectancy theory (Rabey 1979, Vroom 1964) and self-efficacy theory (Bandura 1977, 1982). TAM proposes that two specific belief constructs--perceived usefulness (the extent to which a technological innovation is expected to improve the potential adapter's performance) and perceived ease-of-use (the degree to which the potential adopter expects a technological innovation to be free of effort in use)--are the critical antecedents to an individual's technology adoption decision (Davis 1989).
TAM has seen many applications and extensions in the IS field since its development. Some applications have tested the correlation between innovation usage as predicted by TAM and actual usage behavior (e.g., Szajna 1996), or between self-reports of intended IS usage to actual usage (Straub et al. 1995). TAM has been used to assess the IS adoption characteristics of small firms (Igbaria et al. 1997), to examine adoption differences between genders (Gefen and Straub 1997), and across diverse cultures (Phillips et al. 1994, Straub et al. 1997). Other work has sought to either extend TAM by adding additional constructs to the core model (e.g., Jackson et al. 1997, Taylor and Todd 1995a) or by pursuing a deeper understanding of the two antecedent constructs that predict behavioral intent (e.g., Venkatesh and Davis 1996).
The Perceived Characteristics of Innovating (PCI) Belief Constructs
The PCI belief constructs proposed by Moore and Benbasat (1991) draw on earlier conceptual work by Rogers (1995). Rogers' diffusion of innovations perspective incorporates five innovation characteristics as antecedents to any adoption decision: relative advantage, compatibility, complexity, trialability, and abservability (Rogers 1995). PCI incorporates three of these constructs--relative advantage, compatibility, and trialability--as originally proposed. Relative advantage represents the degree to which an innovation is perceived to be superior to current offerings. Compatibility is the degree to which an innovation meshes with the adapter's current habits and practices. Trialability represents the extent to which a potential adopter believes that the innovation can be adequately tried prior to the adoption decision. Moore and Benbasat (1991) renamed a fourth construct in Rogers' model--complexity--as ease-of-use to be consistent with other emerging models of adoption in the IS literature (e.g., Davis 1989). Ease-of-use represents the degree to which an innovation is perceived to be easy to use.
Four additional constructs are included in the PCI set: visibility, image, result demonstrability, and voluntariness. Arguing that Rogers' abservability construct was not specific enough for usage in IS contexts, Moore and Benbasat (1991) proposed two more distinct constructs in its place: visibility (the degree to which an innovation is visible during its diffusion through a user community) and result demonstrability (the degree to which the benefits and utility of an innovation are readily apparent to the potential adopter). Image--which Rogers originally included as part of the relative advantage construct--represents the degree to which an individual believes that the adoption of an innovation will bestow them with added prestige in their relevant community. Finally, voluntariness reflects the extent to which innovation adoption is perceived to be under the potential adapter's volitional control.
Despite its theoretically rich development and fairly rigorous initial testing, the full set of PCI belief constructs has received relatively little empirical attention (exceptions are Agarwal and Prasad 1997, Gagliardi and Campeau 1995, Moore and Benbasat 1994). IS researchers often cite or discuss...
Read the full article for free courtesy of your local library.
|
Bookmark this article
Print this article
Link to this article
Email this article
Digg It!
Add to del.icio.us
RSS
