| The accumulated empirical literature has consistently shown that computer self-efficacy strongly predicts motivational and behavioral outcomes related to computer technologies. Despite the conceptual importance of the construct, extant computer self-efficacy measures are largely narrowly defined and outdated, having been developed before advances in networked real-time information technologies. As a result, they fail to capture the multidimensional and multifaceted nature of the construct, particularly in more complex technological contexts. The goal of this dissertation research is to develop and establish initial psychometric evidence for a multidimensional measure of computer self-efficacy better suited for use in complex technological contexts. To achieve this goal, I deployed a two-phased research approach.; In phase one, data gathered from an exploratory survey fielded in an organization implementing an Enterprise-wide Resources Planning (ERP) software system was examined. I used part of the survey data to identify factors influencing an individual's confidence in their abilities to use a complex software system. In addition to the survey data, I gathered qualitative information from semi-structured interviews of subject matter experts. From these two sources of data, I created an item pool for a new computer self-efficacy scale.; In phase two, I used a card sorting methodology to refine the item pool. I then developed and administered another survey questionnaire to test the psychometric properties of the new computer self-efficacy scale. The new dimensional computer self-efficacy scale demonstrated high reliability and preliminary evidence for construct validity. |
