Existing models of work group effectiveness have been tested in low-technology manufacturing settings, but not in a high-technology one. Typical features of high-technology environments - substantial automation, computerized scheduling systems, complex production processes, capital-intensive production, high-cost equipment, and high product value - render them significantly different from their low-technology counterparts. The increased use of work groups and teams among high-technology manufacturers raises the question of whether existing models can be generalized to these settings. In this paper, data from 89 groups in the semiconductor manufacturing industry are used to create predictive models of group productivity, job satisfaction, and perceived performance. External variables (such as conflict with supervisors and engineers) are found to be more predictive of productivity, while internal variables (such as conflict among group members) are more predictive of satisfaction. These results highlight the importance of fostering the work group's interaction with external technical support personnel when designing and managing successful work groups in high-technology workplaces.