Effects of computer self-efficacy and system reliability on user interaction with decision support systems

Computer self-efficacy (CSE) has been identified as a major determinant of computer-related ability and usage in organizational contexts. However, there has been little research on the impact of CSE on the use of high level computerized systems such as decision support systems. In order to examine this relationship, participants completed a visual search task with the assistance of an automated decision support system which varied in reliability from moderately reliable (70%) to highly reliable (90%). After completion of the task, a median split was performed on participants’ CSE scores to divide participants into high-CSE and low-CSE groups. We examined the relationship between CSE and trust and utilization of the system. High-CSE participants trusted the system more, complied with it more and generated significantly more hits than low-CSE participants, particularly on trials in which the aid was highly reliable. This indicated that high-CSE levels led to a better ability to gauge the true capabilities of the system. However, all participants uniformly underestimated the true reliability of the system at both levels of CSE. This study provides insight into the potential significance of personality factors in shaping human–technology interaction and has implications for designers of automated systems.