Department of Statistics, UCLA

The concepts and tools developed by Donald Campbell and his colleagues have quite properly dominated the eld of evaluation research. Internal validity, external validity, and later, construct validity and statistical conclusion validity, for instance, are central concepts on which a variety of evaluation research methods depend (Rossi and Freeman, 1993). However, these methods have been developed to understand phenomena that may be empirically observed. In this chapter, we broaden their applicability by considering evaluations undertaken in the virtual worlds manufactured by computer simulations.