Current goal setting studies are working in two directions. The first direction is to identify the cognitive processes of goal setting and the second direction is to examine the moderating effect of task complexity on the goal-performance relationship. The objective of this paper is to integrate the two current directions by examining the moderating effects of two types of task complexity, component complexity and coordinative complexity, on the cognitive processes of goal setting. Two hundred and sixty-seven undergraduates who were taking upper level management courses participated in the study and 226 subjects were included in the final analysis. Subjects performed a stock prices prediction task on computer. Goals were assigned after the subjects have finished the pretest and the subjects were asked to answer questions on their specific self efficacy and performance valence perceptions. Subjects were also asked to set personal goals before performing the test. The cognitive processes models of goal setting for different task groups were compared with the multi-sample analysis of LISREL VII. It was found that subjects who performed the task with low component - low coordinative complexity used a simple motivation mechanism. They motivated their behaviors directly with specific self-efficacy. On the other hand, subjects who performed the more complex tasks used a more rational motivation mechanism that required more cognitive processing. The subjects set personal goals by considering performance valence and specific self efficacy, and they used personal goals to motivate their behaviors. Unlike the simple task for which specific self-efficacy directly affected performance, personal goals were found to mediate the effect of specific self-efficacy on performance for the complex tasks. / Ph. D.