Research practitioners have traditionally collected and analyzed data using experimental designs and inferential statistical procedures. In addition to being difficult to learn and apply, these quantitative procedures generally leave out such variables as attitudes, feelings, and emotions. This paper describes the nature and role of qualitative data collection and analysis procedures in the context of evaluation research or impact assessment. This method is often used by social service professionals to determine if a given program, process, or procedure has worked; focusing only on readily quantifiable outcomes would appear to leave out quite a few essential social and behavioral elements. The document provides a framework for discussing the two complementary aspects of impact assessment, process and outcome, and relative strengths and functions of quantitative versus qualitative procedures. Following this discussion, several commonly applied qualitative procedures are identified, the power of combining numbers and words in the form of multimethod research designs is explained, and their applications are illustrated in a variety of actual evaluation studies. (Contains 23 references.) (LL)