It has recently been determined that generating an explanatory hypothesis to explain a discrepant event is important for students' conceptual change. The purpose of this study is to investigate how students' generate new explanatory hypotheses. To achieve this goal, questions are used to identify students prior ideas related to electromagnetic induction. After showing conflicting phenomena, six college students are asked to suggest explanatory hypotheses to explain the phenomena. Using interviews, the processes of generating explanatory hypotheses are analyzed and three types of hypotheses suggested by students are subsequently identified: theoretical, experiential, and auxiliary hypotheses. In addition, models of generating each type of explanatory hypothesis are also suggested. It is concluded that subjects use similarity-based reasoning to relate their background knowledge or experiences with the conflicting phenomena to be explained. Background knowledge plays a very important role in generating new theoretical explanatory hypotheses. An example of a scientific inquiry activity for improving the skill of "generating scientific hypotheses" is also presented in this paper. (Contains 3 tables and 10 figures.)