Analogical inference is a type of inductive inference that identifies similarities between a source and target domain, and uses those similarities to infer that some further feature of the source domain can be generalized to the target domain. My dissertation aims to help answer two questions about analogical reasoning in science. First, a descriptive question: what role do analogical inferences play in science? I consider this both in the context of traditional human-driven science and contemporary machine learning. Second, a normative question: what normative criteria can be given for determining relevant similarities and to what degree will these criteria be general versus local and context-dependent? I examine current influential normative accounts of analogical inference in philosophy and cognitive science and identify several significant challenges that existing accounts face. I show that there are fruitful criteria that scientists use to make and evaluate analogical inferences that fall outside the scope of these accounts. Finally, I give a positive account that aims to incorporate these additional criteria.