Despite growing diversity in the nursing workforce, there has been limited investigation of the implications of diversity to the performance of nursing workgroups. The aim of this study was to examine relationships among two categories of nursing workgroup diversity (highly visible and job related), intervening workgroup processes (workgroup cohesion and initiative), and workgroup performance (affective and cognitive). A descriptive design with path analysis was used to test two models derived from Pelled's intervening process theory. Registered nurse (RN) and patient data from 222 nursing units at 126 hospitals, collected as part of the Outcomes Research in Nursing Administration Project (ORNA-II), were used. Registered nurses provided demographic data used to estimate workgroup diversity (age, race, unit tenure, educational background, and RN experience) and completed questionnaires measuring workgroup cohesion and initiative. Patients completed a satisfaction survey, and these data, along with unit-level patient falls and medication errors, were used to measure workgroup performance. Our findings failed to support the intervening process theory. No direct effects between diversity and intervening process variables were found. Cohesion and initiative did not mediate the relationships between diversity and workgroup performance. However, workgroup cohesion predicted greater patient satisfaction (p < .01) and higher levels of met expectations for symptom management (p < .01), whereas workgroup initiative predicted fewer patient falls (p < .01). Diversity may have little, if any, effect on nursing workgroup performance. Cohesion and initiative warrant further research as possible mediators of the relationship between patient outcomes and variables like nurse staffing, experience, and expertise.