Πλοήγηση ανά Συγγραφέα "Karaiskos, Dimitris C."

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 1 από 1

A predictive model for the acceptance of pervasive information systems by individuals
Karaiskos, Dimitris C.; Καραϊσκος, Δημήτρης Χ.; Athens University of Economics and Business, Department of Management Science and Technology; Giaglis, George. M.
Pervasive Information Systems constitute an emerging class in the information systems realm motivated by the pervasive (or ubiquitous) computing paradigm. Pervasive computing promises a technological shift away from the desktop computing paradigm towards more ubiquitous forms of computation presence and use. According to Weiser, who envisioned this computing evolution back in 1991, people and environments will be augmented with computational resources that will provide information and services when and where desired in the most acceptable, easy and pleasant way, like a walk in the woods. Today, we observe that this vision gradually becomes a reality through a multitude of pervasive applications taking their position in various real life settings. Nevertheless, the urgency to rush an individual’s world with the latest and greatest pervasive technology must be tempered with an understanding of whether the technology serves appropriately his needs. In other words, this technological shift raises issues regarding the acceptance of pervasive information systems and hence their success. Picking up on this, the work presented in this thesis follows an approach with the aim to correlate pervasive information systems with technology acceptance theories seeking to comprehend the factors that influence the acceptance of these systems. This research effort is further reinforced by the lack of comprehensive inquiries on the phenomenon on behalf of the research community. In fact our, limited, knowledge of pervasive systems acceptance has been garnered through exploratory studies causing this area to lack of evaluation methodologies and metrics (Scholtz and Consolvo 2004). On the contrary, this dissertation aims at building a basis of empirical knowledge with the goal of identifying how the novelty of pervasive information systems, named as pervasiveness, augments the capabilities of technology acceptance theories to predict the acceptance of pervasive systems. To structure the inquiry, a multi-method research strategy is designed and implemented including mostly quantitative research methods. Our research design is backed up by the construct development methodology, which allows us to define and construct an instrument for measuring pervasiveness, and the relevant research literature on technology acceptance, which allows us to formulate the evaluation framework consisting of both acceptance and pervasiveness factors. Starting from the latter objective, an extended literature review revealed to us the modus operandi of technology acceptance theories, the several categories of factors employed in relevant studies along with the limitations existing and the research directions proposed towards more comprehensive models. Aligning the need of the research community to extend technology acceptance theories in that way to acknowledge technology characteristics and our objective to subsume pervasiveness into these theories we formulate our research framework (depicted in Figure 1), which will be rephrased into the research model in a later stage of this study. Knowing how to approach the inquiry of pervasive information systems acceptance we lack of an instrument that appropriately measures the notion of pervasiveness. Considering that our knowledge is limited regarding the pervasiveness factor, and how to measure it, we refer to construct development methodology which consists of sequential steps towards the production of a robust instrument. These steps include initially a thorough literature review for specifying the conceptual domain of pervasiveness. Through this process we defined pervasiveness as the “extent to which an IS consists of interconnected technological artifacts, diffused in their surrounding environment, working together to ubiquitously supports user tasks and objectives in a context aware manner”. Further on we distinguished three founding technological dimensions of pervasiveness, namely ubiquity, diffusion and context awareness and provided statements characterizing them (depicted in Table 1). Based upon the results of the first phase of construct development and particularly the statements of ubiquity, diffusion and context awareness we proceed to the operationalization of these concepts, i.e. the instrument construction phase. During this phase, each statement in the domain is converted into several items in the instrument resulting in the generation of items from and for all the statements that tap each dimension of pervasiveness. Preliminary items are generated following quality criteria, such as length, clarity, and reading efficiency and phrasing according to the selected format of measurement, which in our case is the Likert scale. Successively, the initial instrument of pervasiveness is reviewed by experts in the area for its content validity, that is whether it measures the content of pervasiveness (Straub et al.2004). In doing so, a pre-test, initially, and an expert survey, subsequentially, were held. The pre-test aim is to reveal deficiencies in the initial instrument regarding its format, content, understand ability, terminology, and ease and speed of completion. Responses from the pretestwere collected and adjustments made to the instrument based on the feedback of the respondents. The updated instrument is used next in the expert survey which has the objective to assess each item regarding its relevancy with the concept intended to measure. Experts on pervasive computing (119 in number) were approached from whom 39 responded (a satisfactory response rate of 32%), while 33 answers were ultimately usable after removing incomplete questionnaires. The gathered data were analyzed using the content validity ratio technique (Lawshe 1975) and provided feedback on which items are content valid, i.e. which items are retained or rejected from the instrument. Furthermore, experts’ comments accentuated the shortcoming of using the word diffusion, for one of pervasiveness’ dimensions, as the same word is used for the adoption of information systems and proposed the use of unobtrusiveness as the alternative (a proposition that we adopted). Table 2summarizes the resultant instrument. Moving forward towards the last phase of construct development we aimed in empirically validating the under development construct and to provide, as a result, a valid scale for pervasiveness. Empirical validation is succeeded by two studies, an exploratory and a confirmatory. The exploratory study has the objective to empirically validate the instrument of pervasiveness considering its construct validity and reliability. For this purpose a survey was conducted where 141 participants were asked to fill in a questionnaire comprised by items from the pervasiveness instrument. The participants’ responses were based on the experience gained from a scenario walk through describing the use of a pervasive information system. Then, the gathered responses were analyzed using exploratory factor analysis (EFA)which provided empirical evidence on both construct validity and reliability. In particular, the EFA results further refined the instrument by deleting 12 more items resulting in an instrument of 18 items in total and provided evidence towards defining the structure of the seitems. The confirmatory study has the objective both to endorse the findings of the exploratory study and to further assess the pervasiveness scale. Particularly, through confirmatory factor analysis (CFA) the factors emerged from the exploratory study are tested again providing results over the discriminant and convergent validity, i.e. construct validity. In parallel, factor reliability through multiple measures is also computed. For gathering the empirical data needed for the confirmatory study, a field study was conducted where 128 participants were asked to use a pervasive information system designed and implemented for the purpose of this study and then fill in a questionnaire comprised by the 18 items from the pervasiveness instrument and items capturing a set of acceptance factors. Our findings from CFA and reliability analysis indicate that 16 items are retained and grouped under three distinct factors, as it was initially hypothesized in the conceptual definition of pervasiveness, namely ubiquity, unobtrusiveness and context awareness (see Table 3). Furthermore, construct validity and reliability points out that the specific items indeed measure the concept of pervasiveness, as initially defined. Nevertheless, CFA findings do not provide evidence regarding their causal relationships with the acceptance factors employed in the study. For this reason, a second round of analyses was held, through multiple regression analysis, to reveal the causal relationships of pervasiveness with dominant acceptance factors, or in statistical terms to validate the nomological network of pervasiveness. The relationships tested are depicted in the research model (Figure 2) which was based upon the research framework formulated earlier and operationalized by adopting robust and valid factors from the technology acceptance literature. Our findings revealed the relationships among the factors of pervasiveness and the acceptance factors. In particular, the three dimensions of pervasiveness were found to directly influence the cognitive and affective factors represented by performance expectancy, effort expectancy and perceived enjoyment. On the other hand, facilitating conditions, represented by perceived monetary value, was found to have very weak (or no) relationship with pervasiveness’ factors. Furthermore, very weak or insignificant effects were proved to exist between social factors(Social Influence and Personal Innovativeness) and pervasiveness factors, as correctly hypothesized. Figure 3 summarizes the aforementioned causal linkage of pervasiveness with acceptance factors, i.e. its nomological network. Moreover, we analyzed our data to investigate the mediating effects in our research model. As it has been hypothesized pervasiveness will influence indirectly the dependent variable(Intention) through mediating variables (Performance Expectancy, Effort Expectancy, Perceived Enjoyment and Perceived Monetary Value). Mediation analysis indicates partial support to this hypothesis as Ubiquity was found to be partially mediated by Performance Expectancy and Effort Expectancy, Unobtrusiveness was found to be fully mediated by Effort Expectancy and Perceived Enjoyment while Context Awareness failed to enter the mediation analysis as its correlation with Intention was not adequate. Figure 4 summarizes the indirect effects of pervasiveness on Intention, concluding this way its nomological network. Finally, we evaluated the overall research model by incorporating in the analysis all the variables. The results of multiple regression analysis indicate that Intention to use the pervasive system is predicted by Social Influence, Personal Innovativeness, Perceived Enjoyment, and Effort Expectancy. On the other hand Performance Expectancy and Perceived Monetary Value were found not to influence Intention. Moreover, Ubiquity has been found to influence directly Intention as a result of the partial mediation discussed earlier. Figure 5summarizes both the direct and indirect effects on Intention. Overall, this study offers three important insights that are of value to the research community. The first theoretical contribution of this research lies in outlining the important technological factors of pervasive information systems and in providing an instrument for robustly assessing them. Our research advances existing scholarly work on pervasive information systems by organizing current knowledge regarding the characteristics of this phenomenon under an instrument intended to provide valid quantitative data for further exploitation in technology acceptance studies. As such, our study paves the way for future endeavors within the same thematic area and informs researchers interested in pervasive information systems on their properties while at the same time it equips similar research essays with an instrument capable to enhance their knowledge on the acceptance of pervasive information systems. The second theoretical contribution of this research lies in the identification of the nomological network of pervasiveness. Our findings contribute to the body of literature concerning how pervasive information systems are perceived by its users and how these perceptions influence their intention to accept them. Pervasiveness’ factors were tested along with dominant technology acceptance factors, formulating a robust framework, an inquiry that supported the hypothesized causal relationships between pervasiveness and usage beliefs. In addition, the analysis moved a step forward towards correlating these causal effects with intention, revealing the hierarchy of causal effects that exist from pervasiveness to intention. Finally, it must be accentuated that our contribution is empirically validated, which was not until now the case in the existing research regarding the predicting factors of pervasive information systems acceptance. Last, the third theoretical contribution of this research effort refers to the practice of research. Research efforts towards explaining the interplay of technological characteristics and usage beliefs under the technology acceptance perspective will grow in numbers as the research community considers it imperative. It is our belief that this research contributes in this effort by producing and following a new research approach, backed up by the robust methodology of construct development. As with all research essays, our work have certain limitations that need to be mentioned. The first limitation directly relates to the specific systemic context within which the research has been carried out. The two systems that were utilized for the exploratory and confirmatory study restrain the generalizability of our findings by introducing a certain amount of bias to the resulting observations. Another limitation stems from the fact that the empirical data were collected through the means of snapshot instead of longitudinal research. Snapshot studies impact the investigation of causality among concepts of interest in that relationships are inferred rather than proven. A third limitation is imposed due to sample particularities and characteristics. These limitations cast a certain amount of ambiguity regarding the extent to which our findings are generalizable and transferable to other pervasive information systems and contexts of use. Nevertheless, they open further research opportunities to a plethora of directions, either methodological or theoretical, that could help overcome the deficiencies of the current study. This research implies a plethora of different research directions, targeting either its theoretical anchors or its methodological framing, and as such, it is hoped that it will stimulate supplementary theory and research. First, it would be stimulating to replicate the study with different pervasive systems. This approach would provide the opportunity to explore the effect of system-specific particularities and contextual conditions on the acceptance of pervasive information systems. Second, it would be interesting to adopt an alternative methodological approach to the one followed in this study to investigate the same research questions regarding the acceptance of pervasive information systems. Longitudinal studies or experimental approaches could extrapolate interesting results and weaken that way the effect of common methods bias in the findings and conclusions of this study. Finally, a third direction for future research could involve elaborating on the inherent characteristics of pervasiveness by pointing the theoretical scope to alternative research contexts, such as usability studies or design inquiries.