Computer Use in the Scientific Office

The current research was designed to contribute to the study of scientists' computer use through the development of a linear predictive model of use. Nine variables were selected for study: computer experience, job requirements (profession, time spent on, and importance of 11 work activities), support system size, computer literacy, satisfaction with current tools, overall impact, the importance of computer literacy and skills to management, the importance of computer literacy and skills to colleagues, and interest in computers. There were three components to data collection: (a) on-line recording of participant computer use, (b) a questionnaire, and (c) a discussion period held immediately following completion of the questionnaire. Questionnaire measures were assessed for test-retest reliability and were developed to ensure content validity. Computer use was measured both with a computer-based measurement tool and with the study questionnaire. A cluster analysis on work activity data revealed three distinct clusters of scientists. A single capability with computers factor was derived from a factor analysis on computer literacy, computer experience, and interest in computers. Likewise, a single computer use factor was derived from a factor analysis on measures of use. A linear predictive model of profession + cluster membership + capability with computers accounted for 61 percent of the variance in participant computer use. These findings indicate that while scientists' computer use is determined by a complex set of inter-related variables, a great deal of the variance in scientists' computer use can be predicted from measures of job requirements (viz., profession, cluster) and capability with computers.