This study was primarily an empirical investigation in the field of Information Systems (IS) and the related fields of occupational psychology and management. It focussed specifically on the concept of user satisfaction, the construct of cognitive style as applied to users and systems analysts, and their interrelationships. Prior studies were found rarely to investigate the changes in user satisfaction during system usage. Further, any reference to cognitive style in the IS literature proved to be sparse, open to question and discouraging in terms of its value. By developing and using a new instrument, the System Satisfaction Schedule, or SSS, the present study was able empirically to demonstrate clear patterns of changing user satisfaction during system usage. These were demonstrated, both as a general trend and in terms of its relationship to the cognitive styles of the key players (analyst and user) involved in system development and maintenance. Cognitive style was measured using Kirton's Adaption-innovation Inventory, or KAI. This study was thus able to suggest new rules for system development based on the assessments of the cognitive styles of both users and systems analysts. These rules focussed primarily on simple team choice: which analyst to put with which user. However, inferences for larger system development teams were drawn and suggestions for further research duly made. The present study thus also contributes to the successful practice of system development. To give effect to the above, this study set out to investigate empirically the way user satisfaction changes over 1½ to 2 years of system usage and, as mentioned above, the way user satisfaction is impacted by the cognitive styles of the user and the systems analyst. Most significantly, relationships were studied between user satisfaction and the difference in cognitive style between the analyst and user. It was found that user satisfaction generally rises linearly with usage, and that while the size of the analyst user cognitive differential does negatively impact user satisfaction over most of the time of system use, this effect is only particularly strong for two short periods; one within the first four months of usage and the other in the last three. From these results the new rules for system development mentioned above, followed. In terms of the decline of users' mean perceived severities of individual problems, the exponential decay and reciprocal models were found to fit the data the best. This study developed a new model for the motivation to use, develop or maintain a system (the Mechanical Model), based on its own results and Herzberg's two-factor theory of motivation. In this, Herzberg's hygiene factors have been replaced with the concept of dissatisfiers. These are measured as expressions of dissatisfaction as and when they occur. Their use removes the researcher's need, when designing user satisfaction instruments, to speculate on complete lists of factors which may satisfy users, and which may date as technology and other contextual factors change.