The roles of training, problem representation, and individual differences on performance of both automated (simple) and controlled (complex) process tasks were studied. The following hypotheses were tested: (1) training and cognitive style affect the representation developed; (2) training and cognitive style affect the development and performance of automated processing; (3) training and cognitive style affect controlled process task performance; (4) task representation affects development and performance of automated processes; and (5) task representation affects controlled process task performance. To test these hypotheses, 19 undergraduate students (9 males and 10 females) of varying cognitive abilities were trained in an alphabetic (n=9) or hierarchical (n=10) manner to use a word processor. After training, the subjects' task representation was assessed and they were required to perform both controlled and automatic process tasks. The first hypothesis was not supported; the second, fourth, and fifth hypotheses were supported; and the third hypothesis could not be confirmed. Performance on repetitive tasks associated with automatization was influenced by training style and the mental task representation held by individuals. Task representation was also a significant determinant of performance on complex cognitive-oriented (controlled process) tasks. No effect was found for individual differences. To maximize performance, training and task design should consider the mental representation of the task. Two figures and four tables present study data. (SLD)