Premium
Predicting System Usage from Intention and Past Use: Scale Issues in the Predictors
Author(s) -
Kim Sung S.,
Malhotra Naresh K.
Publication year - 2005
Publication title -
decision sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.238
H-Index - 108
eISSN - 1540-5915
pISSN - 0011-7315
DOI - 10.1111/j.1540-5915.2005.00070.x
Subject(s) - predictive power , variance (accounting) , computer science , measure (data warehouse) , scale (ratio) , explained variation , information system , data mining , machine learning , philosophy , physics , accounting , epistemology , quantum mechanics , electrical engineering , engineering , business
The objective of this study is to provide insights into how the predictive power for computer‐recorded system usage can be improved. Based on 386 responses from actual users of an information system, we examine the predictive power for system usage according to the scales of the predictors used, namely, intention and past use. First, we show that the predictive power of intention can be significantly improved with the choice of an appropriate measure. However, even the desirable intention measure failed to explain two‐thirds of the variance in system usage. Second, the results show that past use as measured by computer‐recorded log data can significantly enhance our ability to predict system usage. Finally, when both intention and past use are controlled for, the explained variance in system usage is shown to vary widely from 20% to 73%, depending on the predictors' scales. Overall, our findings suggest that an accurate prediction of system usage requires a more rigorous approach than that often applied in information systems research.